Fadilla Dissertation

PAGE 36 CONTAIN AIS DATA PROCESSING
KYUSHU
UNIVERSITY
Doctoral Thesis
SAFETY ASSESSMENT FRAMEWORK OF

TRAFFIC SEPARATION SCHEME AS
COLLISION RISK CONTROL OPTION
-PRACTICAL EXAMPLE OF
SUNDA STRAIT IN INDONESIA-
By
Fadilla Indrayuni Prastyasari
Supervised by
Prof. Takeshi SHINODA
Graduate School of Engineering

Department of Urban and Environmental Engineering
Faculty of Engineering
Kyushu University
December 2021
Summary of Dissertation
This paper presents a framework of doing a safety assessment of the newly
implemented traffic separation scheme (TSS) as the risk control option in Sunda
Strait, Indonesia.
The maritime safety and security have become a concern to the International
Maritime Organization (IMO) as an organization that responsible to ensure the
safety and security of the world’s shipping. The good geographical position of
Indonesia which located in the middle of Pacific and Indian Ocean comes with not
only big advantages, but also some challenges. As stated by the United Nations
Convention on the Law of the Sea (UNCLOS) 1982 that an archipelagic state shall
provide free passages for the international route that is later named as the Indonesia
Archipelagic Sea Lanes (IASL). As many vessels are passing the IASL, Indonesia
feels the urgency to set TSS in its water to protect the safety of navigation. As a
result, two traffic separation schemes: Sunda Strait and Lombok Strait were
proposed to IMO and approved in the 6th session of NCSR in January 2019 and
then adopted by the MSC on its 109th session in June at the same year.
The implementation of TSS in Sunda Strait has an objective to reduce the number
of head-on collisions by separating the opposing traffic. TSS Sunda Strait is also
followed by a precautionary area and an inshore traffic zone. However, another
problem is emerged in Sunda Strait related to crossing situations that occurred due
to a heavy traffic nearby the precautionary area as well as in the eastern side of
Sunda Strait. Therefore, a safety assessment is carried out to evaluate the effectivity
of TSS Sunda Strait. This method is recommended by IMO to evaluate an existing
policy or if a new policy is about to be made. The safety assessment consists of five
i
steps that include: 1) hazard identification, 2) risk assessment, 3) risk control
options, 4) cost benefit assessment, and 5) recommendation for decision maker.
This study was started with the hazard identification that utilised AIS data to get
the picture of traffic pattern in Sunda Strait. AIS messages were provided by
AISITS and the data collection process was conducted twice. The first one is from
October 8th, 2018, until October 21 st, 2019, representing the time before setting
TSS. The second one represents a period when TSS is already implemented on April
6th, 2021, until July 24th, 2021. However, in that time interval there are some dates
with data gap, which means no data received. This data gap is excluded from the
analysis during the data processing step. A concept of vessel trips database was
introduced to keep the record of vessel trips that were extracted by means of IALA
Waterway Risk Assessment Program (IWRAP). The distance to closest point of
approach (DCPA) was applied to determine the near-miss situations which the
result was saved in a near-miss situation database.
Probabilistic risk assessment was carried out using time-based analysis with three
point-of-views: crossing zone, course, and vessel type. The total annual frequency
of near-miss crossing situation after the implementation of TSS showed a decrease
compared to before implementing of TSS, which lead to a conclusion about the
effectivity of TSS. However, some risk control options were given to further
minimise the collision risk and maintain the safety navigation in Sunda Strait. Cost
benefit ratio is used to assess the effectivity of TSS in terms of the cost and the
establishment of aids to navigation considered to be cost effective for Crossing
Zones 1,2, and 3. As the last step, recommendations for decision maker were
outlined to support the implementation of RCOs.
Keywords: AIS, Safety Assessment, Traffic Separation Scheme, Risk of Collision,

Near-miss Crossing Situation.
ii
Acknowledgements
This dissertation has finally been fully completed although it is still far from a
perfection. I would like to give my highest gratitude to all supports that I have
received during the process.
Firstly, I am profoundly indebted to my supervisor Prof. Takeshi Shinoda for his

guidance, advice, and inspiration during my study in Japan. Secondly, Prof.
Yoshitaka Furukawa and Assoc. Prof. Takashi Tanaka who have dedicated their
time to be my examiners and gave comments to my research. I would also thank
JICA for making the AUNSEED Net scholarship that giving me a chance to study
in Japan. I am also very grateful to the support from my workplace, ITS Surabaya,
especially Professor Ketut Buda Artana and Dr. Dinariyana who shared their
expertise in risk assessment. As for the AIS data that become an important asset to
my research, I owe my thanks to Mr. Dwi Kristianto and Dwi Suasti.
Live in Japan would be flat without all colleagues that I have had, especially
members of Functional Design of Marine System Laboratory for welcoming me in
a warm environment. It would be hard to survive my three years journey without
supports from Dr. Putu Hangga NP, Dr. Tiago, Liu Guangshuai, Gabriella Bunga
Naen, Theodora Noely, Fernando Lawrens, and Sergio, I am truly lucky to count
you amongst my friends. Special thanks I would also send to M. Akmal Putera
whom with I shared my ideas and obtained great motivation and feedback.
Finally, this work is dedicated to my parents that I eternally grateful for their
continuous and unparalleled love, help and support.
Fadilla Indrayuni Prastyasari
Fukuoka, December 2021
iii
“This page is intentionally left blank”
iv
Table of Contents
Summary of Dissertation ........................................................................................ i
Acknowledgements ................................................................................................ iii
Table of Contents .................................................................................................... v
List of Figures ...................................................................................................... vii
List of Tables ........................................................................................................ ix
Chapter 1 Introduction .................................................................................... 1
1.1 Motivation .............................................................................................. 1
1.2 Aims and Objectives ............................................................................. 3
1.3 Research contributions ......................................................................... 4
1.4 List of Publications ................................................................................ 5
1.5 Outline of the thesis ............................................................................... 6
Chapter 2 State-of-the-Art of Safety Assessment for Vessel Collision ....... 11
2.1 The Concept of Collision Risk Assessment ....................................... 11
2.2 Methodologies for Probabilistic Risk Assessment............................ 14
2.2.1 Ship Domain ..................................................................................... 14
2.2.2 Closest Point of Approach ................................................................ 16
2.2.3 IALA Waterway Risk Assessment Program (IWRAP) .................... 18
2.3 Safety Assessment Framework .......................................................... 19
Chapter 3 Safety Assessment Framework: Hazard Identification ............. 23
3.1 Traffic Pattern in Sunda Strait .......................................................... 23
3.2 Environmental Considerations .......................................................... 30
3.3 Accidents Record ................................................................................. 31
3.4 AIS Messages ....................................................................................... 32
3.5 AIS Data Ingestion .............................................................................. 39
3.6 AIS Data Processing............................................................................ 41
3.7 Vessel Trips Database ......................................................................... 44
3.8 Near-miss Database............................................................................. 49
v
Chapter 4 Safety Assessment Framework: Risk Assessment ....................... 53
4.1 Collision Risk Assessment ................................................................... 53
4.2 Probability Assessment ....................................................................... 54
4.2.1 Crossing zone basis ........................................................................... 55
4.2.2 Course basis ....................................................................................... 56
4.2.3 Vessel type basis ................................................................................ 56
4.2.4 Estimation of the Annual Frequency ................................................. 57
4.3 Risk Acceptance Criteria .................................................................... 57
4.4 Results of Probability Assessment...................................................... 58
4.5 Root Cause Analysis ............................................................................ 60
4.5.1 Bow tie analysis ................................................................................. 62
4.5.2 Safety Barrier for Collision Accidents .............................................. 64
Chapter 5 Safety Assessment Framework: TSS as a Risk Control Option .. 69
5.1 Ships’ Routeing Measures................................................................... 69
5.2 TSS in other countries ......................................................................... 71
5.2.1 TSS Dover Strait ................................................................................ 71
5.2.2 TSS Strait of Malacca and Singapore ................................................ 73
5.3 TSS in Sunda Strait Indonesia ........................................................... 74
5.3.1 Properties of TSS Sunda Strait .......................................................... 76
5.4 Assessing the Effectivity of TSS as an RCO ...................................... 80
5.4.1 Enforcement of COLREGs ................................................................ 86
5.4.2 Priority in Precautionary Area ........................................................... 86
5.4.3 Routeing systems in Eastern Side of Sunda Strait ............................. 87
5.5 Cost Benefit Assessment...................................................................... 87
5.5.1 Cost due to the Implementation of RCO ........................................... 88
5.5.2 Benefits resulted by the Implementation of RCO ............................. 90
5.5.3 Result of Cost Benefit Assessment .................................................... 91
5.6 Recommendation for Decision Maker ............................................... 92
Chapter 6 Conclusion and Future Research ................................................ 93
6.1 Research Findings................................................................................ 93
6.2 Future Research ................................................................................... 95
Reference ....................................................................................................... 97
Appendix A ..................................................................................................... 105
Appendix B ..................................................................................................... 110
Appendix C ..................................................................................................... 115
Appendix D ..................................................................................................... 117
vi
List of Figures
Figure 1.1 Flow of dissertation (part 1). ................................................................. 8
Figure 1.2 Flow of dissertation (part 2) .................................................................. 9
Figure 2.1 Type of encounters from the own vessel point of view. ...................... 11
Figure 2.2 Stages of developing a collision. ......................................................... 13
Figure 2.3 Geometrical collision diameter(Fujii and Shiobara, 1971). ................ 15
Figure 2.4 Ship domains for head-on encounter (left) and overtaking encounter
(right) with distances in cables (Coldwell, 1983). ................................................ 15
Figure 2.5 The definition of CPA during a crossing encounter. ........................... 16
Figure 2.6 Illustration of CPA............................................................................... 18
Figure 2.7 A crossing situation happened after TSS (1). ...................................... 20
Figure 2.8 A crossing situation happened after TSS (2). ...................................... 20
Figure 2.9 Flowchart of the safety assessment framework. .................................. 22
Figure 3.1 Location of archipelagic sea lanes in Indonesia (Dinariyana, 2020). .. 24
Figure 3.2 Traffic pattern in Sunda Strait before TSS. ......................................... 25
Figure 3.3 The nautical chart of Sunda Strait (Indonesian Navy, 2013)............... 31
Figure 3.4 The location of collision accidents in Sunda Strait. ............................ 34
Figure 3.5 Illustration of AIS. ............................................................................... 34
Figure 3.6 The appearance of NMEA message of raw AIS data .......................... 37
Figure 3.7 Flowchart of AIS data handling ........................................................... 39
Figure 3.8 The location of AISITS base stations. ................................................. 40
Figure 3.9 The number of data collected from 2018/10/08 until 2019/10/21. ...... 40
Figure 3.10 Spatial distribution of AIS data. ........................................................ 42
Figure 3.11 Time distribution chart of the highest AIS data received. ................. 43
Figure 3.12 The number of vessels by type at each data set. ................................ 45
vii
Figure 3.13 Workflow to generate a vessel trip database and near-miss database.
............................................................................................................................... 46
Figure 3.14 Visual observation for the trip count .................................................. 46
Figure 3.13 The number of trips that follow each course. ..................................... 47
Figure 3.14 The number of trips made by each vessel type. ................................. 47
Figure 3.15 The number of trips passed by each crossing zone (CZ). .................. 48
Figure 3.16 The number of near-miss on each crossing zone (CZ). ..................... 50
Figure 4.1 Acceptance criteria for individual risk based (UK HSE, no date). ...... 58
Figure 4.2 Basic concept of bow tie diagram (Ehlers et al., 2017). ...................... 63
Figure 4.3 Bow tie analysis for crossing collision in Sunda Strait. ....................... 64
Figure 4.4 VTA of collision between Norgas Cathinka and Bahuga Jaya. ........... 65
Figure 4.5 Safety barriers for collision accidents. ................................................. 66
Figure 5.1 (a) Traffic separation scheme separated by separation zones in 4 and
separation line in 3, (b) inshore traffic zone adjacent with other routeing systems,
(c) precautionary area with a recommended traffic flow (IMO, 2010). ................ 70
Figure 5.2 TSS Dover Strait (Sea-Seek, 2012) ...................................................... 72
Figure 5.3 Number of ships reporting to Klang VTS by type (Jabatan Laut Malaysia,
2021). ..................................................................................................................... 73
Figure 5.4 TSS Strait of Malacca and Singapore (Zhen, 2016). ........................... 74
Figure 5.5 Choke-points, sea lines and mega-hub around Indian and Pacific Ocean
(Rimmer and Lee, 2007). ....................................................................................... 75
Figure 5.6 Properties of routeing measures in Sunda Strait. ................................. 77
Figure 5.7 Heatmap before the implementation of TSS. ....................................... 79
Figure 5.8 Heatmap after the implementation of TSS. .......................................... 80
viii
List of Tables
Table 3-1 Example of vessels that navigating in Sunda Strait .............................. 28
Table 3-2 Record of collision accidents around Sunda Strait during 2010-2019. 33
Table 3-3 Comparison between Class A and Class B AIS (NMEA, 2011). ......... 36
Table 3-4 Information borne by AIS message. ..................................................... 38
Table 3-5 Example of the near-miss database....................................................... 50
Table 4-1 Information about time intervals. ......................................................... 55
Table 4-2 The result of time-based probability assessment for the crossing zone
basis. ...................................................................................................................... 59
Table 4-3 The result of time-based probability assessment for the course basis. 60
Table 4-4 The result of time-based probability assessment for the vessel type basis.
............................................................................................................................... 61
Table 4-5 Details of safety barriers (Part 1) .......................................................... 67
Table 4-6 Details of safety barriers (Part 2) .......................................................... 68
Table 5-1 Coordinates of TSS Sunda Strait. ......................................................... 78
Table 5-2 Information about new time intervals. .................................................. 82
Table 5-3 The result of time-based probability assessment for the crossing zone
basis before and after TSS implementation. ......................................................... 83
Table 5-4 The result of time-based probability assessment for the course basis
before and after TSS implementation. .................................................................. 84
Table 5-5 The result of time-based probability assessment for the vessel type basis
before and after TSS implementation. .................................................................. 85
Table 5-6 Comparison of two cases of the cost assessment.................................. 90
Table 5-7 Comparison of two scenarios of the benefit assessment. ..................... 91
Table 5-8 Results of cost benefit assessment. ....................................................... 91
ix
x
Chapter 1
Introduction
1.1 Motivation
Ship still becomes the backbone for not only energy supply chain but also world
Pentingnya
kapal dan trade. According to the data from International Chamber of Shipping, the shipping
keprihatinannya
terhadap sector transport about 80% of world trade volume and it is approximately 70% of
kecelakaanya
the world trade value. Furthermore, the annual value of the world trade in 2019 had
reached 14 trillion USD (ICS, 2020). Due to this massive world shipping activities,
the safety level of the marine navigation should also be increased or at least kept in
the safe level. In line with this objective, the number of marine accident as well as
its casualties need to be controlled. Although the number of incidents in 2020 shows
slight decrease, several countries are still showing a high number of accidents
within 2015 until 2019, especially for groundings, collisions, and fire or explosion
as stated in the accident reports in Japan, UK, Canada, and Marshall Islands (CTSB,
2021; JTSB, 2021; MAIB, 2021; Maritime Administrator, 2021).
The maritime safety and security have become a concern to the International
Mengapa maritime
safety penting dan Maritime Organization (IMO) as the special agency of the United Nations. This
concern IMO
organization is responsible to ensure the safety and security of the world’s shipping.
terhadapa urusan
maritime safety IMO have established the Maritime Safety Committee (MSC) to deal with all safety
and security matters that cover both passenger vessels and merchant vessels. Along
with that, the Navigation, Communications, and Search and Rescue (NCSR) is a
sub-committee under the MSC that deals with things related to navigation and
1
communication, one of them is the analysis and approval of the ships routeing
measures as well as ship reporting systems (IMO, 2021).
The NCSR holds an annual session since 2014 to discuss about new proposals or
amendments to the existing routeing measures and ship reporting systems. Due to
the high number of vessel collisions at some areas in the past, Germany, France and
sejarah the United Kingdom performed studies about separating the traffic in areas with a
ditemukannya heavy traffic density to reduce the risk of collision and these studies were proposed
TSS
to IMO. Hence, these studies become the first step of separating the opposite traffic
that is now known as the traffic separation scheme (TSS) (IMO, 2010). As per 2010,
there are 157 TSSs that have been adopted by IMO around the globe. IMO adopted
routeing measures are the ones that used by the international shipping. Several
routeing measures are not adopted by IMO, because they are only applied for
domestic shipping. Thus, the regulation is only adopted by the local government
and those are not in the scope of this research. memasukkan sedikit batasan masalah
kaitannya The good geographical position of Indonesia which located in the middle of Pacific
dengan
wilayah di and Indian Ocean comes with not only big advantages, but also some challenges.
Indonesia
dan banyak Indonesia with its approximately 17,500 islands make it the largest archipelago
kecelakaan state and that makes many big cargo ships pass Indonesia water to reach their
kapal
destination. As stated by the United Nations Convention on the Law of the Sea
(UNCLOS) 1982 that an archipelagic state shall provide free passages for the
international route that is later named as the Indonesia Archipelagic Sea Lanes
(IASL) (Yanti, Aviolita and Marsetio, 2020). As many vessels are passing the IASL,
Indonesia feels the urgency to set TSS in its water to protect the safety of navigation.
As a result, two traffic separation schemes: Sunda Strait and Lombok Strait were
proposed to IMO and approved in the 6th session of NCSR in January 2019 and then
adopted by the MSC in their 109th session in June at the same year.
2
1.2 Aims and Objectives
The Government of Indonesia has implemented the traffic separation scheme (TSS)
and other routeing measures in Sunda Strait since July 1st, 2020, to increase the
safety of navigation. The fact that the traffic density in this location is high and it
resulted on several crossing zones observed in Sunda Strait. This research has an
aim to reduce the collision risk in Sunda Strait by means of Traffic Separation
Scheme as a Risk Control Option in a safety assessment framework. To be able to
achieve that research aim, research objectives are organized as follows:
1. Develop a framework for a vessel near-miss crossing situation database in

Sunda Strait by utilizing the available Automatic Identification System (AIS)
data.
This research wants to address the problem of vessel collision due to the high
traffic density in Sunda Strait. That traffic condition leads to a high number of
near-miss in the crossing zones. Hence, a framework of a near-miss crossing
situation database is constructed using the available AIS data.
2. Evaluate the risk of vessel collision in Sunda Strait before and after TSS is
implemented.
The implementation of TSS shall make a positive impact to the safety of

navigation in Sunda Strait. In order to prove that statement, a risk assessment
is performed. A probability assessment is carried out twice, before and after
TSS is implemented. The assessment uses three perspectives: crossing zone,
course, and vessel type basis, to see the problem in a wholistic way.
3. Identify the root causes of the vessel collision by mean of Bow tie analysis.
As the risk level of the collision accident is already known, the root causes of
the vessel collision shall be identified. TSS Sunda Strait has been implemented
only for a short time, thus many aspects could lead to an unsafe trip. The cause
3
is including but not limited to human elements, technical operations, and
policies or regulations.
4. Propose risk control options to decrease the number of vessel collision risk
based on the identified root causes.
The main objective of this research is to provide risk control options to the
existing TSS Sunda Strait, so that the risk of vessel collision can be reduced.
Since Indonesia now has implemented two TSSs in its water, this study is
expected to contribute to the global safety navigation.
5. Assess the cost effectivity of TSS as a risk control option.
To support the result of the risk assessment, the cost benefit assessment is used
to see the ratio of spending due to the implementation of TSS and the benefit
resulted from it. According to the Guidelines of FSA published by IMO, there
are two parameters that can be used to assess the cost-benefit ratio: the gross
cost to averting fatality (GCAF), and the net cost to averting fatality (NCAF).
This research utilises the GCAF to avoid any biases during the calculation of
economic benefits.
6. Give recommendations to the Government of Indonesia to improve the safety

of navigation in Sunda Strait.
As the final step of the safety assessment framework, recommendations are

given to the Government of Indonesia which are based on the result of the
assessment.
1.3 Research contributions
This background of this research is due to the implementation of the traffic

separation scheme in Sunda Strait, which is the passage for both ocean-going cargo
vessels and roro / passenger ferries. As the research aim and objectives have been
4
explained in the previous sub-chapters, the main contributions given by this
research are:
1. Introducing a concept of vessel trip database and vessel near-miss crossing

situation database for the vessels navigating in Sunda Strait that can help the
authority to map the traffic pattern based on the AIS data,
2. Evaluating the effectivity of TSS Sunda Strait providing the framework of
safety assessment as recommended by IMO,
3. Contradicting the level of risk in Sunda Strait before and after the
implementation of TSS by the mean of a probability assessment for crossing
collision that includes three bases (i.e., crossing zone, course, and vessel type)
which is in line with IMO Instruments Implementation Code (III Code) about
the obligation of coastal state,
4. Demonstrating a framework of a bow tie analysis for finding the relation of the
cause-consequence to understand the root cause of the vessel collision in TSS
Sunda Strait which is supported by the safety barrier that can prevent a vessel
collision to take place, and
5. Providing a recommendation and address the limitation of the existing
regulations and policies in TSS Sunda Strait as a reference to set additional
regulations to support TSS.
1.4 List of Publications
Those contributions explained in the previous part is discussed and written in

following publications:
1. Fadilla Indrayuni Prastyasari and Takeshi Shinoda (2019): Near-miss Analysis

of Marine Traffic in Sunda Strait Indonesia, The 4th International Conference
on Marine Technology (SENTA), 2019, Surabaya, Indonesia.
2. Fadilla Indrayuni Prastyasari and Takeshi Shinoda (2020): Ship’s Collision
Frequency Assessment for Introducing Traffic Separation Scheme in Sunda
5
Strait in Indonesia, 日本航海学会誌 NAVIGATION / 211 巻 pp. 17-18,
https://doi.org/10.18949/jinnavi.211.0_17.
3. Fadilla Indrayuni Prastyasari and Takeshi Shinoda (2020): Near-miss
Detection for Encountering Ships in Sunda Strait, IOP Conference Series:
Earth and Environmental Science, Volume 557, 2nd Maritime Safety
International Conference (MASTIC) 18-Jul-20, Surabaya, Indonesia,
doi:10.1088/1755-1315/557/1/012039.
4. N. D. Wuryaningrum, D. W. Handani, Fadilla Indrayuni Prastyasari and A. A.
B. Dinariyana (2020): Frequency Analysis of Ship Collision and Its Impact on
The Fulfillment of Supporting Facilities and Route Changes Due to
Implementation of Sunda Strait TSS, IOP Conference Series: Earth and
Environmental Science, Volume 557, 2nd Maritime Safety International
Conference (MASTIC) 18-Jul-20, Surabaya, Indonesia, doi:10.1088/1755-
1315/557/1/012042.
5. Fadilla Indrayuni Prastyasari and Takeshi Shinoda (2021): Probability
Assessment of Crossing Situation in Sunda Strait, ICSOT (International
Conference on Ship and Offshore Technology) Politeknik Perkapalan Negeri
Surabaya (PPNS) and Royal Institute of Naval Architecture (RINA) (will be
published in RINA transactions).
6. Fadilla Indrayuni Prastyasari and Takeshi Shinoda (2021): Study on
Probabilistic Risk Assessment Model for Crossing Situation in Sunda Strait,
The 6th International Conference on Marine Technology (SENTA) 2021,
Surabaya, Indonesia, (will be published in IOP Conference Series).
1.5 Outline of the thesis
The dissertation is written based on the flow as illustrated in Figure 1.1 and Figure
1.2 and it can be described as follows.
Chapter 2 provides a comprehensive literature study about the state-of-the-art of

safety assessment, especially for vessel collision. As vessel collision has become
one of the major accidents in the maritime sector in several countries. This chapter
6
is divided into three sub-chapters, the first one is about the concept of risk analysis
in general. The second sub-chapter discusses about the risk analysis practice in the
vessel collision cases. Next, the concept of a safety assessment framework is
proposed. There are five steps for safety assessment that is utilised in this research,
those are: hazard identification, risk assessment, risk control options, cost-benefit
assessment, and lastly recommendations. The hazard identification process is
outlined in Chapter 3, risk assessment is discussed in Chapter 4, and lastly in
Chapter 5 the risk control option, cost benefit assessment and recommendations are
examined.
Chapter 3 describes in detail about the hazard identification. As the marine traffic
of Sunda Strait becomes the concern of this research, so the traffic pattern is
extracted here to find the hazard. Following that, the AIS data handling is also
explained as a basis to construct a near-miss situation database.
Chapter 4 explains about the process of risk assessment. As the result of AIS data
handling process, the near-miss database that has been developed is used as the
basis for calculating the probability of near-miss situations. In addition to the
probability assessment, the annual frequency is also assessed to understand the level
of the risk in Sunda Strait that is based on the number of near-misses per year.
Besides the quantitative analysis of the probabilistic risk assessment, this chapter
also observes the root causes of collision accident. The bow tie analysis is used and
is supported with the safety barriers analysis. The purpose is to understand the
causes of the accident as well as the barriers that could prevent it.
Chapter 5 discusses about the traffic separation scheme (TSS) as a risk control
options (RCO) for reducing the risk of collision in Sunda Strait. To prove the
effectivity of TSS, the probabilistic risk assessment is performed once again in this
chapter. The cost benefit assessment is also conducted to know the effectivity of
TSS in terms of economic value. Lastly, the recommendations to the Government
of Indonesia to maintain the safety of navigation in Sunda Strait.
7
Finally, Chapter 6 presents the research findings, conclusions, and
recommendations for further research.
Figure 1.1 Flow of dissertation (part 1).
8
Figure 1.2 Flow of dissertation (part 2)
9
.
10
Chapter 2
State-of-the-Art of Safety Assessment
for Vessel Collision
2.1 The Concept of Collision Risk Assessment
Marine traffic accidents are often associated with collisions or grounding. But in a
more specific scope, ship collisions get more attention because there are many
factors involved in the incident, such as faulty manoeuvring, environmental
conditions, inevitable causes, etc., not to mention the high casualties caused by this
accident that reached 26.2% between 2011 to 2018 according to the European
Maritime Safety Agency (EMSA) (European Maritime Safety Agency, 2019). A
study was performed to analyse traffic accidents in Japan, and it discovered that a
high risk of collision was found in locations nearby the port or in a narrow channel
like a strait, so more attention is needed to be given in this area (Fujii and Shiobara,
1971).
Figure 2.1 Type of encounters from the own vessel point of view.
11
During the ships’ operation, they must ever be encountering other ships. This
condition is thought as the beginning of the vessel collision. Action to be taken
when the ship is experiencing this situation is described in the International
Regulations for Preventing Collisions at Sea (COLREGs) Section II: Conduct of
Vessels in Sight of One Another Rules 13-15 and each of the situation can be
illustrated by Figure 2.1. When the own vessel is encountering target vessels, there
are two possible actions that can be taken, namely give way and stand on, as stated
in Rules 16 and 17 depending on the position in which they are experiencing.
COLREGs also stipulates in Rule 10 about how ships should act when they are
navigating in TSS.
As previously explained, the encountering situation is the initial stage of a collision

accident. Figure 2.2 explains how a collision can occur. Although there is no exact
guidance on how this stage is defined, but this could be a rough definition about
each condition that would be experienced before the collision take place (Hilgert,
1983). In the initial stage, the ship which was not originally involved in the risk of
collision, then faces a situation where the bearing between two vessels is not
changing enough and if vessels keep going in that course and speed, they will meet
(collide) at a certain point. The next stage is where the vessels dangerously
approaching each other and the action of one vessel alone may not be enough to
avoid risk of collision. We may assume that the vessel is in a close quarter situation
if one vessel is entering into the others turning circle. If both vessels cannot perform
an immediate action to avoid the collision, an immediate danger situation is
developed where the action of give-way vessel alone is not enough and both vessels
shall make the best coordinated immediate action to avoid the collision. Failed to
perform this coordinated immediate action, a collision accident could immediately
occur and no further action to avoid could be done.
Risk assessment is a useful tool to help in determining the condition of the area of
interest and decide any actions that will be taken to ensure the safety of navigation.
The result of risk assessment is beneficial for every party responsible for the safety
of ship's navigation as well as for Vessel Traffic Service (VTS) operators, which
12
Figure 2.2 Stages of developing a collision.
have a responsibility to supervising the traffic in ports, inland waterways, or coastal

areas (Silveira et al., 2021). As mandated by IMO Instruments Implementation
Code (III Code) through IMO Resolution A.1070(28) that coastal states have
obligations to “periodically evaluate its performance in respect of exercising its
rights and meeting its obligations under the applicable international instruments”
and it could be achieved through a systematic risk assessment. Since the
implementation of TSS is currently in its second year, there is a need for serious
monitoring followed by the regulation enforcement to keep the routeing systems
relevant with its aims.
Despite of the importance of risk assessment, neither The International Regulations

for Preventing Collisions at Sea (COLREGs) nor International Convention for the
Safety of Life at Sea (SOLAS) have a comprehensive description about the
definition of risk of collision and guidelines about how to do the collision risk
assessment. Nevertheless, several studies have been attempted to find out the
13
perceived risk of collision due to the marine traffic (Chin and Debnath, 2009) and
tried to propose a way to avoid the collision by means of communication with local
VTS (Baldauf and Wiersma, 1998).
Several methodologies have been developed to quantify the risk of collision, for
example: ship domain, closest point of approach (CPA), and the IALA Waterway
Risk Assessment Program (IWRAP), that briefly elaborated in the following
paragraphs.
2.2 Methodologies for Probabilistic Risk Assessment
2.2.1 Ship Domain

Fujii and Shiobara started a study about the traffic analysis in 1971 and expressed
that the number of collisions 𝑁col in time interval 𝑇 inside the waterway area 𝑆,
where both ships are sailing on a random course is related to the geometrical
collision diameter and can be notated as:
(1)
Where 𝜌1 and 𝜌2 are the traffic density of each ship that can be estimated with the
Equation 2, P is the probability factor, D is the geometrical collision diameter as
drawn in Figure 2.3, and 𝑉1 and 𝑉2 represent the velocity of each ship.
(2)
𝑄 is the observed traffic volume, 𝑊 is the width of the channel, and 𝑉 is the
velocity of a ship.
A noticeable study has also conducted to further develop about the quantification
of the risk of collision. According to this study, a safe navigation could be achieved
by providing a certain safe space surrounding the ship that is known as a ship
domain. It can be defined as “the effective area around a ship which a navigator
would like to keep free with respect to other ships and stationary objects” (Goodwin,
1975). She did not suggest any distance for the ship domain but stated that the shape
14
of the domain would be asymmetrical due to the regulation in COLREGs and
divided the domain into three sectors: starboard, port, and astern.
A further development of the ship domain conducted to be applied in restricted

waters, to address the difference in the study area between Fujii that focused on the
restricted water and Goodwin that concerned about the open sea (Coldwell, 1983).
Two domains for merchant ships were proposed in the study that accommodate
head-on and overtaking situations that featured in Figure 2.4.
Figure 2.3 Geometrical collision diameter(Fujii and Shiobara, 1971).
Figure 2.4 Ship domains for head-on encounter (left)

and overtaking encounter (right) with distances in cables (Coldwell, 1983).
15
2.2.2 Closest Point of Approach
The development of marine radar with automatic radar plotting aid (ARPA) could
create an automatic calculation for target vessels course, speed, as well as closest
point of approach (CPA). Combination of ARPA and AIS could broaden the
coverage to detect the target vessels as well as track the movement of smaller
vessels at the same time (Lin and Huang, 2006). On the other hands, many collisions
experienced due to the lack of skill to assess the situation, as the ARPA and AIS
can only give the calculation for CPA at the present moment (Learnmarine, 2018).
The CPA value calculated by ARPA not always showing a right value as stated by
IMO MSC.64(67) about performance standards for radar equipment that this value
could have an error that approximately could be up to 0.3 nautical mile (nm). A
different value of CPA could be resulted by the AIS device, depending on the
sensors’ errors of the vessel.
The CPA, as shown in Figure 2.5, pointed out that an own vessel is at its closest
location to the target vessel when t = t3. This indicator has both spatial and temporal
features, namely the distance to the closest point of approach (DCPA) and the time
to the closest point of approach (TCPA), respectively (van Iperen, 2015). The CPA
approach is a good methodology to be used in a non-accident probabilistic risk
assessment approach as no real accident is involved in the calculation. The concept
of CPA is used in this study to determine a vessel encounter to have a near-miss
crossing situation or not.
Figure 2.5 The definition of CPA during a crossing encounter.
16
Many researchers have proposed the value of CPA in different conditions. Research
conducted by Goodwin, for example, stated that the safe DCPA is 2.35 nm for the
open sea (Goodwin, 1975). Meanwhile, Park in his research found that the DCPA
is 0.15 nm and the TCPA is 3 minutes for Korean waters (Park, 2014). Furthermore,
research by Fukuto and Imazu proposed the value of safe DCPA as 1.0 nm and the
safe TCPA is 5 minutes (Fukuto and Imazu, 2013).
This study evaluates the DCPA of passing and crossing vessels in Sunda Strait over
time during the time intervals previously mentioned. Among those DCPAs that
have been proposed by the previous researches, an interview with a deck officer of
a cargo ship was conducted to get the idea about what the crew feels about each
DCPA. The result of this interview is that the DCPA from Fukuto and Imazu
(Fukuto and Imazu, 2013) with the distance of 1.0 nm is taken for categorizing the
vessel encounter condition, considering the condition of Sunda Strait that has a
passing and crossing traffic. If two vessels are encountering each other and the
closest distance between each other is less than 1.0 nm, the encounter is categorized
as a near-miss crossing situation. The general formula of the DCPA calculation is
involving the position and speed of both own vessel and target vessel and the
equation is outlined as follows (Lenart, 2017):
(3)
where (4)
is the relative speed, and are relative speed components of the target vessel
in x and y-axis, and all symbol and notation are depicted in Figure 2.6.
The calculation about CPA is often mistakenly done, as it involved the position of
the ship, so a reference point must be known beforehand. Several research take the
midship as the reference point, while the CPA should be calculated based on
position of the AIS device/antenna. As ship is a large structure that has a length
around 100 meters or more and if we mistakenly calculate the CPA point, we could
17
Figure 2.6 Illustration of CPA.
underestimate the risk of collision, which is undesired in this situation.
2.2.3 IALA Waterway Risk Assessment Program (IWRAP)

Another study is conducted to calculate the collision frequency in a form of
software called The BaSSy ToolBox (GRISK) and this is done by a combination of
several institutions, namely Technical University of Denmark, GateHouse
(Denmark), SSPA (Sweden), and VTT (Finland), which is funded in part by The
Danish Maritime Foundation and Det Nordiske Ministerråd, and this software is
simplified to be IALA Waterway Risk Assessment Program (IWRAP). IWRAP is
a tool to calculate the frequency of collision and grounding that is endorsed by IMO
to support the newly proposed policy that it can reduce the risk of collision and
grounding.
IWRAP introduces this following equation to find the frequency of collision, 𝜆𝑐𝑜𝑙
(Friis-Hansen, 2008):
𝜆𝑐𝑜𝑙 = 𝑃𝐶 × 𝑁𝐺 (5)
18
Where 𝑃𝐶 is a causation factor and 𝑁𝐺 is the geometric number of collision
candidates. The value of 𝑃𝐶 in Equation (5) can be calculated by using fault tree or
event tree analyses or even with a more complex approach that can utilize the
conditional probability between each variable such as Bayesian Network model.
The causation factor can be related to the human errors, not only those who operate
ships but also those who are in charge of maintaining ship traffic around that area.
As it is reported by Squire (2003) that human factor is the first cause of collisions
and any hazardous incidents in the sea, followed by sophisticated navigational
methods and equipment and overcrowding area. The number of causation factor or
𝑃𝐶 that is used in this paper is 0.5 × 10−4 for head-on collision, 1.1 × 10−4 for
overtaking collision, and 1.3 × 10−4 for crossing collision as suggested by IWRAP
to be the default value, because there is no further analysis made to calculate the
value of the causation factor. In addition to the causation factor, 𝑃𝐶 , the value of
geometric number, 𝑁𝐺 , in Equation (5) will depend on the type of collision that is
divided as head-on, overtaking, crossing, merging, and bending collision.
In this study, the DCPA calculation is conducted by means of the IWRAP. This
software offers a great help in analysing the traffic as it can perform a traffic
simulation as well as CPA calculation as well. In addition to the DCPA calculation
PAHAMI
MAKSUDNYA results, IWRAP also provides the sign for indicating encounters that have the DCPA
value equals to or less than the setting value. Figure 2.7 and Figure 2.8 illustrate the
near-miss crossing situation in Sunda Strait on July 7th, 2021 and July 3rd, 2021.
The DCPA of the first situation is 0.53 nm and the second situation is 0.29 nm.
2.3 Safety Assessment Framework
Safety assessment is a complete method that is used to understand the safety level
of a system, which followed with the recommendation on how to improve or retain
the safety level. IMO as an organisation that deals with the safety related matters of
the marine sector in the world also endorses Formal Safety Assessment (FSA) to be
used as a mean to evaluate new regulations or to compare existing regulations with
19
Figure 2.7 A crossing situation happened after TSS (1).
Figure 2.8 A crossing situation happened after TSS (2).
the newly proposed ones (IMO, 2018b). So far, there have been several studies
conducted on safety assessment, particularly using the FSA as recommended by
IMO, for policies in the maritime world. For example, the usage of FSA which
assessed the risk produced by the marine traffic in Istanbul Strait (Görçün and
Burak, 2015). Another example of FSA application is for container ships that
discussed about the accident that ever happened to this type of ships and concluded
that there are still a wide room for improvement related to the safety aspect of the
container ship (Wang and Foinikis, 2001).
20
Although FSA has been used in several studies and is also recommended by IMO
to be used in evaluating policies related to the maritime sector, this method also has
several drawbacks. The lack of transparency is considered as one of them, which is
a contradiction of the purpose of this method that wants to give a transparency
during the assessment. By using the same input, the results can be different because
of the expert judgment factor that largely determines the final result of FSA
(Kontovas and Psaraftis, 2009). In addition, this method requires a cost benefit
assessment where the economic value of each risk control option is calculated and
compared with the benefits that can be obtained. The calculation of the economic
value is not so easy to do because not all aspects in the assessment can be quantified,
especially in terms of cost. Although there are already methods to calculate the cost-
effectiveness in relation to safety of life, such as Gross Cost of Averting a Fatality
(Gross CAF) and Net Cost of Averting a Fatality (Net CAF), in fact they are not
practical to apply.
Based on the methodology, safety assessment can be categorised into qualitative

assessment, such as fish-bone analysis (Rahim and Sunaryo, 2019), Hazard and
Operability study (HAZOP) (Fadhil, Handani and Dinariyana, 2020), etc. While the
other category is quantitative assessment which involves numerical calculations or
mathematical models by performing statistical analysis of historical data or what is
commonly called as probabilistic risk assessment, such as fault tree and event tree
analysis (Cem Kuzu, Akyuz and Arslan, 2019; Sakar et al., 2021), etc.
This study proposes the concept of a safety assessment framework that is carried
out to a traffic separation scheme as a measure to reduce the risk of collision in
Sunda Strait by combining qualitative and quantitative assessment methods to get
more comprehensive results. The steps taken are following the flowchart shown in
Figure 2.9 that includes: 1) hazard identification, which is utilising a qualitative
method to define the hazard existed in Sunda Strait; 2) collision risk assessment,
which focus to the probability of a near-miss situation; 3) risk control options; 4)
cost benefit analysis; and 5) recommendation for decision maker.
21
Figure 2.9 Flowchart of the safety assessment framework.
22
Chapter 3
Safety Assessment Framework:
Hazard Identification
In the process of performing a safety assessment, the first step that must be taken is
to identify the hazard that can endanger the safety of a system. Hazard itself is
initially defined as “potential unsafe factors that may lead to personnel injuries or
financial losses” (Hammer, 1972). As the definition is getting broader, the purpose
of doing the hazard identification remains the same which is to identify all possible
events that can be treated as a threat to the existing system and prioritise which of
those that can be discarded or taken to the further analysis.
This study observed the traffic pattern in Sunda Strait to understand where the
hazard is located by extracting the traffic pattern into courses and crossing zones.
To be able to assess the hazard further, Automatic Identification System (AIS) data
is utilised. Some processes are undertaken to extract the information out of it and
use it to develop the near-miss situation database.
3.1 Traffic Pattern in Sunda Strait
Sunda Strait is located between Java and Sumatera islands. Both islands are
connected by a mean of roro/passenger ferries that is in operation between Port of
Merak and Port of Bakauheni round trip. The ports are run by the state-owned
company named ASDP Indonesia Ferry. Most of the vessels are also owned by this
company, although several of them are owned by private companies that have many
roro ferries all around Indonesia. The distance from Port of Merak to Port of
23
Bakauheni approximately is 16.5 nautical miles and it takes around two hours by
regular ferries and one hour by express ferries.
In addition to the roro/passenger ferries, Sunda Strait is also passed by the ocean-
going cargo vessels. This situation happens due to the responsibility bounded to
Indonesia to provide a free passage, as Indonesia is an archipelagic state that has
ratified UNCLOS 1982. Following this condition, Indonesia has set three
archipelagic sea lanes (IASLs) through the Government Regulation Number 37 and
Number 38 of 2002 that later is amended by Government Regulation Number 37 of
2008 (Schofield and Arsana, 2009). The location of each IASL is drawn in the
Figure 3.1. As seen on the figure, all those IASLs span in the north-south direction.
The IASL 1 is located in the western part of Indonesia, started from Sunda Strait
and going north until South China Sea. The IASL 2 is situated on the Lombok Strait
up to Makassar Strait. While IASL 3 has a more complicated scope in the south part
which includes Banda Sea and Arafura Sea up to the north until reach the Molucca
Sea (Melchior and Baumert, 2014). Two main channels in Sunda Strait have been
outlined in previous paragraphs. In addition to those channels, domestic vessels also
found navigating in this strait. One channel is found parallel to the IASL 1 and
located in the eastern side of Sunda Strait close to the most western part of Java
Island. Most of the vessels that navigate in this course are coming from or to the
IASL 1 IASL 3
IASL 2
Figure 3.1 Location of archipelagic sea lanes in Indonesia (Dinariyana, 2020).
24
industrial area around the Cilegon which is close to the Port of Merak. In this
location, there are several commercial ports that managed and ran by IPC and
several companies, for example Candra Asri Petrochemicals, Asahimas Chemicals,
Bakrie, etc. (JICA, 2003).
Another course is also observed in Sunda Strait. However, this one has a different
direction if compared to other courses that passed by cargo vessels. This channel is
going from the southwest part to the northeast part of Sunda Strait. Domestic cargo
vessels are dominating this passage as the direction is rather going to the north part
or Java Island. To sum up, at least four major channels are spotted in Sunda Strait
and to understand the location of every one of them, Figure 3.2 gives the illustration
about the traffic pattern before TSS is implemented. The density plot is given in
this figure to show the traffic intensity of each course in Sunda Strait. In addition,
each course is named with C1 until C8. The letter C indicates the course and the
number following it is the number of the course.
The characteristics of each course in Sunda Strait is explained as follows.
Figure 3.2 Traffic pattern in Sunda Strait before TSS.
25
1. Course 1 and Course 2 (C1 and C2)
These courses represent the IASL 1 that has been dedicated as the course for
cargo vessels that are going to pass through Indonesian territory. These courses
are dominated by vessels engaged in a trip from Australia, Europe, or West
Asia going to East Asia (Vietnam, Japan, etc.) and most of them are general
cargo and bulk carriers. Vessels engaged in a route from the Hindia Ocean to
the East Asia or in the northbound direction is classed as C1. While C2 is the
opposite of C1 which has a southbound direction. These courses are important
passageways in Sunda Strait as big vessels are using both courses alternately.
The other important courses in Sunda Strait are Course 3 and Course 4, which
accommodate roro/passenger ferries that sail from Port of Merak to Port of
Bakauheni and round trip. Based on the actual schedule published by the Port
of Merak, there are 50 trips per day commenced from the Merak, which make
the traffic in Sunda Strait vicinity becomes very crowded. Course 3 is from
Port of Merak going in western direction and Course 4 is on the opposite
direction from Port of Bakauheni to the eastern direction. Due to the Ministerial
Regulation of the Ministry of Transportation No. 88 of 2014, the vessel that
can give a service in Merak and Bakauheni is vessels with the capacity of 5,000
gross tonnage (GT) and above (Ministry of Transportation, 2014). This
regulation also gives a restriction to the maximum number of vessels that can
be operated here to only 68 vessels.
Although these two courses have a parallel direction with the Course 1 and
Course 2, a heavier and diverge traffic is observed in the eastern part of Sunda
Strait. Similar with C1 and C2, these courses are dominated by merchant
vessels. Most of the vessels are sailing in the domestic route, from the Java Sea
to the industrial ports close to the Port of Merak. Before the implementation of
TSS, several vessels engaged in an international voyage can also be found in
these courses. TSS implementation does not affect the traffic pattern in C5 and
C6 as there is no routeing measure set in this area.
26
These courses are passed by domestic cargo vessels with smaller dimension
sailing from the west part of Sumatera to the Java Island and return. Both
courses are spanned diagonally and intersect with Course 1, Course 2, Course
3, and Course 4. Bulk carrier, tanker and general cargo are still found to be the
most frequently passing vessel in these courses. Course 7 is going to the
northeast direction, while Course 8 is going to the southwest direction.
Apart from those eight courses that have been explained previously, it is also
important to look at the crucial positions nearby the Port. Vessels engaged in
an anchoring position is categorised as C9, while vessels positioned in a port is
labelled with C10, to make the analysis easier. This labelling is performed to
accommodate the vessel operation that includes not only sailing, but also
anchoring and berthing or at port position. These situations are essentials and
need to be put into account because it would contribute to the traffic condition
in Sunda Strait. Several anchorage zones that found around this area are (IPC,
no date):
a. Anchorage area for Panamax and Cape Size ship located in the area
bounded by (6°00'30" S, 105°55'00" E), (6°00'30" S, 105°54'00" E),
(6°01'30" S, 105°54'00" E), and (6°01'30" S, 105°55'00" E).
b. Anchorage area for ships that will be berthing on Jetty Indoprin, Redeco,
BMT, etc., is in area bounded by (5°53'42" S, 106°00'12" E), (5°52'16" S,
106°00'15" E), (5°53'00" S, 105°59'30" E), and (5°53'00" S, 106°00'57" E).
c. Anchorage area for ship reparation is on the area bounded by (5°57'00" S,
106°08'00" E), (5°57'00" S, 106°10'00" E), (5°58'30" S, 106°08'00" E), and
(5°58'30" S, 106°10'00" E).
d. Anchorage area for roro/passenger ferries that serve the route from Merak
to Bakauheni and vice versa, is provided in an area bordered by (5°57'20"
S, 105°58'45" E), (5°56'24" S, 105°59'06" E), (5°56'37" S, 105°58'57" E),
and (5°56'42" S, 105°59'18" E).
27
A few examples of the vessels that navigate through Sunda Strait during the period
of October 8-14, 2018, are shown in Table 3-1. From this table, we can understand
in which course they were sailing at along with the type of vessel, its dimension, as
well as its port of origin and destination.
Furthermore, as shown in the Figure 3.2 that the vessel traffic in Sunda Strait makes
several crossing zones where two or more courses intersect each other in an area.
The risk of collision could be higher in crossing zones because there is a possibility
for a vessel to encounter another vessel in head-on, overtake, crossing, merging,
and bending situation. Seven crossing zones (CZ) are observed here and the
characters of each one of them are briefly described in the following paragraph.
Table 3-1 Example of vessels that navigating in Sunda Strait
Course Type of Vessel Length in Port of Origin Port of Destination

meter [Country] [Country]
1 General cargo 275 Jeddah [SA] Port Klang [MY]
2 Bulk carrier 292 Newcastle [AU] Puerto Drummond [CO]
2 Bulk carrier 189 Kwinana [AU] Jeddah [SA]
2 General cargo 127 Bangkok [TH] Mokpo [KR]
2 Other vessel 169 Higashiharima [JP] Napier [NZ]
3 Roro ferries 114 Merak [ID] Bakauheni [ID]
4 Roro ferries 134 Bakauheni [ID] Merak [ID]
4 Roro ferries 151 Bakauheni [ID] Merak [ID]
5 Bulk carrier 292 Newcastle [AU] Zhanjiang [CN]
5 Bulk carrier 189 Suralaya [ID] Pangkalan Susu [ID]
6 Oil products 105 Gresik [ID] Ciwandan [ID]
tanker
6 General cargo 121 Sakai [JP] Jakarta [ID]
7 Bulk carrier 224 Suralaya [ID] Muara Pantai [ID]
7 Bulk carrier 149 Tarahan [ID] Suralaya [ID]
8 Gas tanker 160 Tanjung Uban [ID] Kwinana [AU]
8 Bulk carrier 146 Bojonegara [ID] Ciwandan [ID]
28
1. Crossing zones 1, 2, and 3 (CZ 1, CZ 2, and CZ 3)
The CZ 1, CZ 2, and CZ 3 are in the west area of Sunda Strait. The CZ 1 is an
intersection between the course that follow IASL 1 (Courses 1 and 2) with
Course 7 and Course 8 which follow a diagonal course, and with Courses 3 and
4 which designated for roro/passenger ferries. It is expected that crossing zone
1 would have the highest number of crossing situation, due to this crossing
zone is made by 6 courses that intersect each other. The crossing zone 2 is the
intersection between Courses 1 and 2, the passage for cargo vessels, and Course
3 and Course 4, the courses for roro/passenger ferries. Hence, crossing situation
is expected between ocean going cargo vessel and roro ferries. The crossing
zone 3 (CZ 3) is located a little bit to the east side of the CZ 2, which is the
potential location of crossing situation made by Courses 7 and 8, which
dominated by domestic cargo vessels with Courses 3 and 4 for roro/passenger
ferries.
2. Crossing zones 4 and 5 (CZ 4 and CZ 5)
The CZ 4 and CZ 5 are crossing zones that involving Course 5 and 6 dominated
by cargo vessels and Courses 3 and 4 for roro/ferries. Both locations are in the
east area of Sunda Strait close to the Port of Merak. The CZ 4 is made between
Courses 5 and 6 with Course 3, which accommodating the roro/passenger
ferries going from Port of Merak to the Port of Bakauheni. While the CZ 5 is
also involving Courses 5 and 6 together with Course 4, which is the passage
for crossing roro/passenger ferries going from Port of Bakauheni to the Port of
Merak.
3. Crossing zones 6 and 7 (CZ 6 and CZ 7)
These crossing zones are made due to the extension of Course 5 and Course 6
that meet with Course 7 and Course 8. As domestic cargo vessels are frequently
pass the Courses 7 and 8 that going to or from the eastern part of Indonesia, so
the vessel engaged in those courses might encounter a crossing situation in CZ
5 or CZ 6. The difference is, the vessel engaged in Course 7 and Course 8
would meet vessels engaged in Course 5 and Course 6 that sail to the north
direction in CZ 6, whereas in CZ 7 the vessels sailing in Course 7 and Course
29
8 might encounter with vessels sailing in Course 5 and Course 6 that are going
to the eastern direction.
In addition to the vessel courses and crossing zones, the vessels that sail in Sunda
Strait are also categorised based on their type. This channel accommodates twelve
types of vessels: bulk carrier, chemical tanker, container, crude oil tanker, fishing
vessel, gas tanker, general cargo, oil products tanker, roro / passenger ferry,
pleasure vessel, support vessel, and other vessel. We could find any type of vessels
in all courses, except Course 3 and Course 4 that are dedicated for the
roro/passenger ferries.
3.2 Environmental Considerations
Sunda Strait connects the Indian Ocean and the Java Sea. Its seabed condition has
a variation of depth as it has angular sub-basin as well as steep slopes due to the
fault movement that make the bathymetry in this area becomes more complicated.
An active volcano named Anak Krakatoa is also situated in the middle south of
Sunda Strait that recently in 2018 had an eruption and led to a lethal Tsunami that
killed more than 400 people (Heidarzadeh et al., 2020). This makes the GoI
designated a dangerous area in the south part of Sunda Strait that is bounded by
Krakatau Island, Panaitan Island, dan Sertung Island, due to the risk for eruptions
that can occur at any time. The position of the Krakatau Island in Sunda Strait can
be observed in the Figure 3.3.
Sunda Strait is also home to several corals and reefs as shown in the Figure 3.3
above, one of which is the Terumbu Koliot. It is a coral reef located in 5°55'0" S,
105°49'1" E that is approximately 1 nm from the IASL 1 axis and on the northwest
side of Sangiang Island in Sunda Strait. Another one is the Terumbu Gosal which
located in 5°50'25" S, 105°53'31" E or at the northern part of Sunda Strait that also
in north of the Sangiang Island. The presence of coral reefs around Sunda Strait
could potentially create a risk of grounding incidents for vessels passing through
IASL 1 from the Indian Ocean to the Java Sea or vice versa. In addition, according
30
Figure 3.3 The nautical chart of Sunda Strait (Indonesian Navy, 2013).
to the Decree of the Minister of Forestry of the Republic of Indonesia No. 55 of

1993 which established Sangiang Island as a Marine Tourism Park area, made this
area protected from ships so as not to disturb the biodiversity in this park.
3.3 Accidents Record
As a high traffic is observed around Sunda Strait and those vessels sail on different
courses, the probability of a collision accident to occur should be high too. Back in
the days, several accidents and incidents took place in this location as a result of
those conditions. There are at least 10 accidents that consist of collisions, allisions,
or in-contact happened during 2010 until 2019 around Sunda Strait that were
experienced by both cargo vessels and/or roro or passenger ferries. One of the
example is a collision between Bahuga Jaya, a roro/passenger ferry, and Norgas
Cathinka, a Singaporean registered gas tanker, that happened in 2012 (NTSC, 2013).
This accident resulted on seven fatalities and severe damages for both ships.
Moreover, grounding incidents have also taken place not less frequent compared to
the previous type of accident. The major grounding incident was experienced by a
31
container ship “Hanjin Aqua” in 2015 due to stuck in the Koliot reef (NTSC, 2016).
This vessel sailed from Port of Adelaide in Australia to Jakarta, Indonesia while
carrying around 2,300 TEUs in her cargo hold. It is written in the report that the
incident happened due to the several roro ferries convoyed during their operation in
crossing Sunda Strait and made the ship master misjudged the situation.
However, regardless the number of accidents and incidents in Sunda Strait, the
official report is only found for few of them, while the rest are only reported in a
news article. The scarcity of data and reports have always become a massive
problem in the maritime sector. Especially in Indonesia, where many stakeholders
that are taking care of the maritime safety and security, such as BAKAMLA
(Indonesia Coast Guard) an independent bureau, KPLP (Indonesia Sea and Coast
Guard) which is under the Ministry of Transportation, and POLAIR (Marine Police)
as a sub-ordinate of the Indonesia National Police (CSIS, 2020), which make the
regulation more complicated.
Table 3-2 is featuring several collision accidents in Sunda Strait that can be
recorded by the author (NTSC, 2013, 2016, 2018, 2020) and the location is shown
in Figure 3.4. The table shows that the ship collision is often involving both
roro/passenger ferries and any type of cargo vessels. As previously mentioned,
crossing zones in Sunda Strait could be formed due to crossing traffics originated
from two or more courses. The record of the collision in this table gives a proof that
those crossing zones possess a higher chance of collision compared to the rest area
in Sunda Strait. Additionally, several accidents are also situated close to the port,
as the density here is also high due to anchoring activities, while many ships are
found to be in manoeuvring mode as well.
3.4 AIS Messages
Automatic identification system (AIS) is a broadcasting system that enable a vessel

to be identified by other vessels and shore-based stations. It can be used for tracking
the vessel movements as the data is continuously send by the AIS transponder
onboard as presented in Figure 3.5. This device shall be installed onboard a vessel
32
Table 3-2 Record of collision accidents around Sunda Strait during 2010-2019.
No. Date Ships Involved Location Losses

1 2019/12/03 Roro/passenger ferry Port of Bakauheni Not reported
Roro/passenger ferry
2 2019/10/30 Oil tanker Suralaya water, Merak Hull damage
Oil tanker
3 2019/04/22 Roro/passenger ferry 1.5 nm from Port of 1 fatality and 2
Roro/passenger ferry Merak injured
4 2017/08/27 Roro/passenger ferry Suralaya water, Merak Vessel
Roro/passenger ferry structure
Roro/passenger ferry damaged
5 2017/07/05 Roro/passenger ferry Port of Bakauheni Not reported
6 2014/10/01 Roro/passenger ferry Sunda Strait (not Not reported
Cargo ship mentioned precisely)
7 2014/05/03 Roro/passenger ferry 3 nm from Port of 24 vehicles
Cargo ship Bakauheni damaged
8 2014/01/28 Roro/passenger ferry Sunda Strait (not Not reported
Chemical tanker mentioned precisely)
9 2012/09/26 Roro/passenger ferry 4 nm east if Rimau 7 fatalities
Gas tanker Balak Island
10 2010/05/19 Roro/passenger ferry Teluk Jakarta Not reported
Cargo ship
with 300 GT or above that engaged in an international voyage and 500 GT or above
that is engaged in a non-international voyage. While for the passenger vessel, the
AIS transponder shall be installed onboard regardless the size (IMO, no date).
Back in 1989, a grounding accident was experienced by an oil tanker “Exxon

Valdez” in Alaska made the oil carried gushed away and polluted the water due to
the damaged hulls. This accident is remembered as the most catastrophic oil spill
in the U.S. back in that time which made the congress gave an order to the coast
33
Figure 3.4 The location of collision accidents in Sunda Strait.
Figure 3.5 Illustration of AIS.
guard to develop a system that can be used as a ship tracking system. Initially, AIS
was created as a mean to broaden the navigating capability of the seafarers onboard
34
a ship, as the visual observation, VHF radio, and radar still have several
shortcomings. On the other hand, they desired a digital and autonomous system that
can be less prone to the human error (Cutlip, 2017).
As a follow up for this occurrence, IMO later in 1998 made a resolution MSC.
74(69), Annex 3 that recommending the use of AIS which is useful for improving
the safety of navigation, protecting the marine environment, and streamlining the
operations of vessel traffic services (VTS) by reducing the risk of collisions
between ships, managing information about the ship, and as a mean of
communication between ship-to-shore (IMO, 1998). Then in 2000, to enforce the
usage of IMO adopted Chapter V Regulation 19 of International Safety of Life at
Sea (SOLAS) and then required the carriage of AIS started from 2004 for newly
built ships as well as for existing ships with certain specifications as written in this
regulation.
There are two types of AIS devices recommended by IMO to be installed onboard
vessels, Class A and Class B. Class A AIS is the ordinary AIS transponder for
merchant ships and passenger ships, while Class B AIS is installed in the smaller
vessels that have less space and require less cost to build. The development of Class
B AIS has a background to ensure the safety of smaller ships as they often cannot
be seen by larger vessels, especially for fishing vessels that are made from non-
metal material which is undiscoverable by radar. While in fact, a lot of collisions in
Indonesia are involving cargo vessels and smaller vessels, as experienced by a
tanker named MT. Gloria Sentosa and a 18-meter traditional cargo carrier Jukung
Irpansya back in 2011 (NTSC, 2012). The features of each AIS class are mentioned
in Table 3-3 Comparison between Class A and Class B AIS (NMEA, 2011).
Despite some advantages offered by AIS, several considerations should be given to

address the problems of AIS. For example, the coverage of AIS antenna is limited
to 15-20 nautical miles when installed onboard (Last, Hering-Bertram and Linsen,
2015). It could be wider if it is placed in a higher and unrestricted place, as in a base
station. However, base station is not always available in any place that makes the
coverage of the AIS is rather hard to handle. In addition, the AIS message could be
35
Table 3-3 Comparison between Class A and Class B AIS (NMEA, 2011).
Features Class A Class B

Installed on Vessels engaged in an Smaller vessels
international voyage, and it
is mandated by IMO which
size is above 300 GT or all
passenger vessels
Transmit power 12.5 watts (nominal), 2 watts
2 watts (low power)
Frequency range 156.025 – 162.025 MHz @ 156.025 – 162.025 MHz @
12.5/25 KHz, DSC 25 KHz, DSC (156.525
(156.525 MHz) is required MHz) and 12.5 KHz are
optional
Unique SOTDMA (self-organizing CSTDMA (carrier-sense
communication amongst Class A’s) polite to Class A’s)
access scheme
a problem too. An incomplete AIS data often happens due to either human error
(not inputting the information correctly) or technical problem such as, not enough
characters or error during encrypting or decrypting the message (Bošnjak,
Šimunović and Kavran, 2012).
The AIS data is transmitted through an encoded message that follows National
Marine Electronics Association (NMEA) 0183 and 2000 protocols. The example of
raw AIS data can be found in Figure 3.6, which contains information that cannot be
easily readable if the messages have not been decrypted yet. The most important
part of the message lies in the end because this part consists of static and dynamic
data that can be useful for getting the ships’ information. The payload data is
decoded by using ASCII payload armouring to be 6-bit binary number, then all 6-
bit numbers are concatenated to be a binary payload message. This payload message
is the one that needs to be chunked to extract the information from it. For example,
36
Figure 3.6 The appearance of NMEA message of raw AIS data
MMSI number is located in bit 8 for 30 bits, the SOG is on bit 50 for 10 bits, heading
is on bit 128 for 9 bits, and so on. A brief description about information borne by
AIS as outlined in IMO A 29/Res.1106 is given in Table 3-4.
Due to all benefit and drawback of the AIS, numerous studies have been utilising
this system to help the vessels’ crews to increase their awareness during operating
the vessel. Most of them are about collision avoidance (Miyake, Fukuto and
Hasegawa, 2015). AIS is very powerful to be used in this research because it has
information about the ships’ position which is sent regularly, autonomously and in
almost real time. In addition, AIS is also employed to determine marine spatial
planning (Le Tixerant et al., 2018), estimating atmospheric emission (Coello et al.,
2015), and tracing illegal oil discharge from vessels (Liu et al., 2021).
In this study, the AIS data is also used to get the information of vessels that passing
Sunda Strait to be able to map its traffic pattern. Since the purpose of this research
is to conduct a safety assessment on Sunda Strait TSS, the AIS data must go through
several processes before it can be analysed using the proposed framework. The
steps taken in AIS data handling can be seen in Figure 3.7 which includes: 1) raw
AIS data ingestion, 2) data processing that consists of data profiling and data quality
handling, 3) data storage in the form of vessel trip database, and 4) analytics ready
data to be a near-miss situation database.
37
Table 3-4 Information borne by AIS message.
Static Data Dynamic Data Voyage-related Safety-related

Data Data
Maritime Timestamp in Ship’s draught Short safety-
Mobile Service coordinated universal related message
Identity (MMSI) time (UTC)
Name and call Ship’s position in Type of
sign longitude and latitude hazardous cargo
(if any)
IMO number Course over ground Destination and
(COG) ETA
Length and Speed over ground Route plan
breadth (SOG)
Type of ship Heading
Location of Navigational status
electronic
position fixing
system (EPFS)
Rate of turn (ROT)
38
Figure 3.7 Flowchart of AIS data handling
3.5 AIS Data Ingestion
Data ingestion can also be interpreted as data collection, where AIS data is taken
from the data provider which in this case is AISITS. It is a start-up company in the
field of technology, especially AIS and it has wide coverage in Indonesia. In 2017,
AISITS installed a base station on the west coast of Java Island at Point A which is
around Sunda Strait and it had a capability of receiving AIS data from ships with a
coverage of 40-50 km, approximately. However, in its operation, instability was
often experienced in terms of data reception. Then the base station with the same
coverage was moved to Point B which is in Serang City in 2020. To get a clear
illustration of the location of the base station, see Figure 3.8.
Raw AIS data with the extension of .NMEA is fetched from the server. Figure 3.9
shows the number of AIS data received by the base station every day during
October8th, 2018 until October 21st, 2019. The bar chart here is showing that the
AIS base station did not receive any data on several dates. AIS data is received very
well in the very first week during that period (October 8th, 2018 until October 14th,
2018). Another lump of data is also received during March 2019 until April 2019,
39
Figure 3.8 The location of AISITS base stations.
40,000
35,000
30,000
The number of data
25,000
20,000
15,000
10,000
5,000
Date
Figure 3.9 The number of data collected from 2018/10/08 until 2019/10/21.
although during this period the number of data is not as much as the previous time
interval. The next lump of data is spotted on April 13th, 2019 until April 19, 2019.
This time, the number of data is as much as the data in October 2018. As provided
in the graph that the rest of the date only few data are observed, this limitation has
40
to be addressed in the future development of this study. It can be solved by getting
more data from another provider or by expanding the coverage of the AIS antenna.
The data described above is a data ingestion process carried out before the
implementation of TSS Sunda Strait. In addition to the dates presented above, data
ingestion was also carried out for the second time after the implementation of TSS
Sunda Strait which was carried out on July 1st, 2020. The data downloaded in this
second batch was carried out over a period of approximately 4 months, April 6th,
2021, until July 24th, 2021. The data obtained from this second batch is much higher
than the first batch which reached four times the highest amount of data taken in
the first batch. This could possibly happen because the GoI enforces the usage of
AIS through the Ministrial Regulation of the Ministry of Transportation no. 7 of
2019 that is implemented on August 20th, 2019, regarding the obligation to install
and activate AIS devices for Indonesian-flagged vessels as well as foreign-flagged
vessels sailing in Indonesian waters (Ministry of Transportation, 2019).
3.6 AIS Data Processing
The data ingestion is followed by the decrypting process which is done through a
Python module named “pyais”. The ready data is saved in a comma separated file
(.csv) which is equipped with information such as timestamp, MMSI, ships’ name,
dimension, position, heading, COG, SOG, ROT, etc. The ready data is plotted in
the map to see the distribution spatially by means of IWRAP. Figure 3.10 depicts
the spatial distribution of AIS data on March 2nd, 2019. During this period, most of
the data received by the base station lie in the northern part of West Java Province
(yellow area), while the focus area is in Sunda Strait (red area). The focus area is
defined as a rectangle bounded by four points: P1 (-5.78018, 105.631), P2 (-5.78018,
106.071), P3 (-6.07616, 106.071), and P3 (-6.07616, 105.631).
The data processing step comprises of two parts, data profiling and data quality
handling. Data profiling is a process of examining, analysing, and creating a useful
summary from the data. In this study, time distribution charts are generated to better
41
Figure 3.10 Spatial distribution of AIS data.
visualise the spread of AIS data within a day. In this report, the data from one week
each from 2018, 2019, and 2021 are taken as dates that have the highest amount of
AIS data in that period and notated as data set 1, data set 2, and data set 3,
respectively. As no data was obtained in 2020, this can be an improvement for the
further study to get data for that period to further complete the analysis on each year.
Figure 3.11 displays the distribution of AIS in one day. On the horizontal axis there
is time in the 24-hour period, while on the vertical axis there are dates in 3-week
intervals. On the right side of the distribution chart, information regarding the
amount of AIS data received by the base station are shown, followed by the total
number of ships, the number of trips made by those ships, the number of ferries and
lastly the number of trips made by roro/passenger ferries. As explained in the
previous chapter, Sunda Strait hosts not only cargo vessels but also roro/passenger
ferries. Therefore, it is important to include the number of these vessels on this chart.
By looking at this chart, it can be seen that in several dates, especially those before
the implementation of TSS, there were a lot of data gaps where no data was received.
Missing data or data gap is a problem that is often encountered and can affect the
results of the analysis. In theory, there are two methods that can be used
42
Figure 3.11 Time distribution chart of the highest AIS data received.
to overcome this data gap problem, namely data omission and data imputation (Nieh
et al., 2014). The data omission method removes samples that have missing data
from the next analysis. This method is easy to perform, however it reduces the
amount of sample noticeably. As for the imputation data, statistical analysis is used
to fill in the missing data. The data entered can be in the form of mean, median, or
other more advanced methods.
In this research, data omission is used to handle missing data in the received AIS
data set. Data omission is done by taking the dates that have the largest amount of
data in a certain period of time and omitting the rest of the data. This is done because
in one day, the time period in which missing data appears is very wide, so that if
43
data imputation is carried out, the results obtained will not be too accurate. However,
this imputation data can be carried out in further studies, to find out how the results
that can be obtained using this method are compared to the results of using the data
omission method.
The result that is obtained from the AIS data processing is a summary of the number
of ships divided by type on the selected dates, as shown in Figure 3.12. Data set 1
(October 8th – 14th, 2018) and Data set 2 (April 13th – 19th, 2019) show that the
highest number of vessels are roro/passenger ferries, but a different result is
generated by Data set 3 (July 1st – 7th, 2021) which indicates that the support vessel
is the ship with the highest number during a week. As for the second and third
position, all three data sets show the same results, those are oil product tankers and
bulk carriers, respectively. The results of Data set 1 and Data set 2 only represents
the traffic characteristic of large vessels, because it only received the data from the
Class A AIS transponder, which installed in the vessel above 300 GT. Smaller
vessels like conventional fishing vessels do not transmit Class A AIS data to the
base station, which make them undiscoverable in these time intervals. However,
during the Data set 3, several fishing vessels and pleasure vessels are discoverable.
This difference could occur due to the implementation of a Ministerial Regulation
that enforces the installation and usage of AIS.
3.7 Vessel Trips Database
After the AIS data is processed and sorted, the next process is to process the data
into a vessel trips database. It is necessary to group the vessel movements to make
one continuous track, as one AIS datum alone only represents one certain point at
a certain time. The process of making the vessel trips database follows the flow
chart depicted in Figure 3.13. This figure shows that the initial step is importing the
AIS data that have been processed previously to the software named IALA
Waterway Risk Assessment Program (IWRAP) to make a density plot. From there,
the trip data are extracted including trip count, course number, and crossing zone
number, as well as information about the vessel. The trip count contains the number
44
Figure 3.12 The number of vessels by type at each data set.
of trips made during the length of each data set. From the number of trips counted,
each trip is assigned to a course number. However, one trip can be labelled with
more than one crossing zone, because a vessel course may have more than one
crossing zone along the way. Thus, all these information are stored for each unique
MMSI that represent a particular vessel.
In the process of counting the vessel trips generated by each vessel, the need to
validate the result of the IWRAP is emerged. As this software has an ability to
connect all coordinates of a vessel, but this feature also has a time constraint (i.e.,
when the time between two coordinates is exceeding the time limitation, the trip is
separated into two different trips). Hence, a visual observation is required to
validate it by comparing the result of trip count resulted by IWRAP with the number
of trip that is visually observed. An example as shown by Figure 3.14 presents that
the number of trips extracted by IWRAP are two.
However, when the date and time of each trip are carefully observed, it is concluded
that those two trips actually are one same trip that disconnected one from another.
Hence, the number of trips from the IWRAP is fixed into the right one. In addition
to that reason, trips made by roro/passenger ferries that is generated by IWRAP can
be validated with the daily schedule released by the operator.
45
Figure 3.13 Workflow to generate a vessel trip database and near-miss database.
Figure 3.14 Visual observation for the trip count
Based on the traffic data that has been obtained for Sunda Strait area and has been
described as a density plot as shown by Figure 3.2, seven crossing zones that can
be potential hazards for the traffic in Sunda Strait are witnessed. Based on the
available AIS data, the number of trips that pass through each course on each data
set are as illustrated in Figure 3.13. This graph shows that Course 3 and Course 4
46
have the highest number of trips with extremely small differences because both are
courses for roro/passenger ferries that cross Sunda Strait daily. The same results are
produced by all data set which means that most of the vessel traffic in Sunda Strait
is contributed by the crossing courses. Next, in the second place is traffic that have
an operation on the east side of Sunda Strait, namely Course 5 and Course 6.
Following them, Course 8 and Course 9 are in third place, where these two courses
represent the condition of the vessel which is not sailing due to at anchor position
or berthing at the port. Lastly, the traffic of Course 1 and Course 2 share the
minority of the heavy traffic in Sunda Strait during all data sets.
Figure 3.15 The number of trips that follow each course.
Figure 3.16 The number of trips made by each vessel type.
47
If viewed from the perspective of the vessel type, the roro/passenger ferries indeed
have the highest trips in Sunda Strait over time. According to the Land
Transportation Management Center (BPTD Provinsi Banten), the number of trips
made by the ferry per day was 100 trips in 2018, so that in 7 days the trips produced
could reach 700 trips. This number is matched with the AIS data received for
Course 3 and Course 4 in 2018. Oil product tankers and bulk carriers are on the
second and third position during the data set 1 and data set 2. Whilst the data set 3
yields a difference result that shows support vessels on the second place and oil
product tankers on the third place.
According to the Figure 3.17, the number of trips that pass through each crossing
zone has increased every year, especially in 2021 where the increase was
experienced significantly. The three data sets show the same results for which
crossing zone is crossed the most each year, namely CZ 4. However, different
results observed for the second and third highest number of trips that pass through
the crossing zone. The data set collected before the setting of TSS, both data set 1
and data set 2 nominate CZ 5 and CZ 3, while the data set collected after the setting
of TSS appoint CZ 2 and CZ 1 as the second and third in rank, respectively.
Figure 3.17 The number of trips passed by each crossing zone (CZ).
48
3.8 Near-miss Database
The DCPA calculation is conducted by means of the IWRAP. This software offers
a great help in analysing the traffic as it can perform a traffic simulation as well as
CPA calculation as well. In addition to the DCPA calculation results, IWRAP also
provides the sign for indicating encounters that have the DCPA value equals to or
less than the setting value. The crossing situation analysis is performed in all
crossing zones mentioned in Chapter 2.2.2 by setting the value of the DCPA with
the distance that has been chosen by the deck officer as 1.0 nm. This value is chosen
based on a consideration that the width of Sunda Strait is only 16 nm, so it is
categorised as a narrow channel. When the value of DCPA is set to be bigger than
1.0 nm, knowing that the vessel is navigating in a narrow channel, it is hard to
perform a manoeuvre with a large distance with another ship. However, any action
to avoid the risk of collision shall be taken before the ship reaches its DCPA. near-
miss. Although the second and third highest are not the same between the number
of trips and the number of near-miss, but the differences are not too significance.
Table 3-5 shows an example of a near-miss database. The information recorded into
this database includes the date, time, in which crossing zone does the near-miss
happen, the MMSI number of the two vessels involved, the type of ships, the course
where the ship made a trip, heading, COG, average speed, and the role of each ship
based on the COLREGs Rule 15 regarding crossing situations. All information
contained in this database is then utilised for the analysis in the next step.
Furthermore, as a summary from the near miss database of all time intervals, Figure
3.18 displays the number of near-miss occurred on each crossing zone. The result
shows a good agreement with the result of the trip database, which has CZ 4 as the
crossing zone where the near-miss took place the most. But looking at the next rank,
the CZ 5 and CZ 1 has the second and third highest number for the data set 3, just
the same as data set 2. As for the data set 1, the second highest are CZ 3, while CZ
5 and CZ 1 has the same number of occurrences in the third highest. From this
situation, it is concluded from both graphs that the number of near-miss crossing
49
situation is related to the number of trips in each crossing zone. This is supported
by the fact that both CZ 4 has the highest number of trips as well as the number of
near-miss. Although the second and third highest are not the same between the
number of trips and the number of near-miss, but the differences are not too
significance.
Table 3-5 Example of the near-miss database
50
Figure 3.18 The number of near-miss on each crossing zone (CZ).
The conditions presented above are taken into consideration before deciding that
the heavy traffic conditions followed by the large number of near-miss crossing
situations at each crossing zone constitute a hazard in Sunda Strait. Thus, the seven
crossing zones (i.e., CZ 1 – CZ 7) are defined as potential hazards in Sunda Strait.
51
52
Chapter 4
Risk Assessment
4.1 Collision Risk Assessment
According to Kaplan and Garrick (1981), the risk analysis is expected to answer
three questions: 1) what can go wrong?, 2) what is the likelihood of that happening?,
and 3) what are the consequences? A concept of risk assessment by Kristiansen
(2013) is used in this paper to study the risk level of the marine traffic in Sunda
Strait before and after the implementation of TSS, can be written as follows:
𝑅 = 𝑃×𝐶 (6)
In this theory, the risk (R) is affected by both probability (P) and the consequences
(C). However, the term of frequency is usually used more frequently to replace the
probability to represent the number of the event in an interval of time (i.e., the
number of near-miss crossing situations per year) (Čorić et al., 2021). The
guidelines for Formal Safety Assessment (FSA) released by IMO also explains
about the definition of probability. IMO divides probability into two, frequentistic
and Bayesian. The first one can also be defined as a relative frequency that can be
stated as the ratio of events to the total number of possible events. While the second
one is defined as the degree of confidence in the occurrence of an event that can be
measured by the scale of 0 until 1 (IMO, 2018b).
As explained before, the second aspect of a risk is the consequence which represents
the effect or the outcome of an accident. In the case of a vessel collision,
53
consequences could be in the form of structural damage to the ships’ hull that could
lead to flooding and affect the stability of the ship (Zhang et al., 2021). Another
example of collision that could generate a structural damage can also be
experienced when the collision is involving not only vessel but also another object,
for example, an offshore platform (Mujeeb-Ahmed and Paik, 2021). Furthermore,
a case of collision accident could also lead to an environment damage such as oil
spill that can cost not only money but also cause a massive pollution to the sea (Lu
et al., 2020).
The following paragraphs below explain the framework used for assessing the risk
of vessel collision accidents due to a dense traffic in TSS Sunda Strait. Probability
assessment carried out by considering the occurrence of near-miss crossing
situations in each crossing zone.
4.2 Probability Assessment
Although Equation 6 is a broadly accepted formula for determining the risk, the
probabilistic risk assessment has been discussed more than the consequences. The
previous researches (Macduff, 1974; Liu and Shi, 2020) have discussed thoroughly
about a formula to calculate the collision frequency that can be written as:
(7)
The frequency, 𝑓, is affected by the number of vessels in conflict or in another term

can be said as near-miss, N and the probability that the vessel cannot perform an
evasive manoeuvre in a given time or known as the causation factor, 𝑝𝑐 .
This study proposed a new perspective to perform a risk assessment that is used to
evaluate the risk of collision in Sunda Strait before the implementation of TSS
which is based on the probability. The near-miss situations are used as the basis for
performing the probabilistic risk assessment and then followed by the calculation
of annual frequency to understand the level of the risk.
An approach on how to do the probabilistic assessment is attempted, which is based

on the time basis, or specifically 24 hours basis. Due to the AIS data that have been
54
collected have many data gaps which make the size of data ready data become small,
the analysis is only performed on two time-intervals. Each of which has a one-week
duration. The information about each time interval is shown in Table 4-1. As written
in the table, the vessel counts and trip counts of both time intervals are similar,
hence a comparison can be made using these two data sets.
The probability and frequency of near-miss crossing situations in this study are
assessed by using three bases, namely crossing zone basis, course basis, and vessel
type basis. The database, which has been constructed from the collection of near-
miss crossing situation containing information about the crossing zone, course, as
well as the vessel type, is gone through a calculation to get the total number of near-
miss crossing situation for each basis. The equation used for calculating the
probability of each basis is a simple mathematical model that is explained below.
4.2.1 Crossing zone basis

The definition of the near-miss crossing situation probability in the crossing zone
basis or is how likely a trip that crosses would experience a near-miss situation
and the following equation is used to computing it.
Table 4-1 Information about time intervals.
Data set no. Time Interval 1 Time Interval 2

Start date and time October 8th, 2018 April 13th, 2019
01:16:17 GMT 00:00:18 GMT
End date and time October 14th, 2018 April 19th, 2019
04:44:48 GMT 09:35:30
Time duration 106 hours 150 hours
Vessel counts 295 301
Trip counts 1221 1550
55
(8)
Where is the number of near-miss situations in and is the number of
trips in . The number of 𝑛𝐶𝑍𝑖 in Equation 8 is multiplied by 2, because one near-

miss situation in a crossing zone is made by two trips of two vessels.
4.2.2 Course basis
The near-miss crossing situation probability under the course basis or , can be
defined as the likelihood of a trip in course to engaged in a near-miss situation
in that course and the value of is calculated by using Equation 9.
(9)
Where and are the number of near-miss crossing situations and the number
of trips in each course, respectively.
4.2.3 Vessel type basis

Lastly, the near-miss crossing situation probability under the vessel type basis or
can be defined as the chance of a trip made by a type of vessel to have a near-
miss situation. The equation for calculating this probability is outlined as follow.
(10)
Where and are the number of near-miss situation and number of trips in ,
respectively. Similar to the calculation in the Equation 9, the calculation does not
need to be doubled for the as two vessel type is already counted in one near-
miss crossing situation.
56
4.2.4 Estimation of the Annual Frequency
Beside the probability, it is also important to know the annual frequency of the
collision based on the near-miss situations calculated above. Due to the scarcity of
the collision data in Indonesia, the annual frequency is estimated based on the
number of trips as well as the number of near-miss crossing situations. Firstly, the
number of all possible trips in a year, must be estimated as a foundation on doing
the frequency assessment.
(11)
The definition of annual frequency that is being used in this study is the ratio of
near-miss crossing situations to the total possible trips in a year. Probability differs
from the annual frequency because the probability directly takes the ratio between
the number of near-miss crossing situations and the total trips, while in the annual
frequency the time dimension is introduced. As the data used in this study is a
weekly data, so it is necessary to multiply the result with 52 to annualise the result.
The calculation of the annual frequency for every basis are performed by using
these following equations:
(12)
(13)
(14)
4.3 Risk Acceptance Criteria
The annual frequency of the near-miss crossing situations in Sunda Strait shall be
categorized under the acceptance criteria to understand whether TSS is improving
the frequency or make it even worse. It is supported by the International Maritime
Organization Maritime Safety Committee (IMO MSC) 72/16 about the Formal
57
Safety Assessment (FSA) decision parameters including risk acceptance criteria
(Norway delegates, no date). Hence, an acceptance criteria from the UK HSE (UK
HSE, no date) for individual risk is used to categorized the level of frequency of
near-miss crossing situations and it is shown in Figure 4.1. The figure separates the
reversed triangle into three parts: unacceptable, As Low as Reasonably Practicable
(ALARP), and acceptable. The situation falls under the unacceptable level when
the frequency is more than 1.00E-03 per year. Right below the unacceptable is the
ALARP, when the frequency of 1.00E-03 until 1.00E-06. If the frequency falls
under this condition, some efforts shall be taken to maintain the risk level or even
further reduce it. The effort mentioned here will be discussed thoroughly in the risk
control option part in Chapter 5.3. The last one is the acceptable level that can be
achieved when the frequency is less than 1.00E-06.
4.4 Results of Probability Assessment
Initially, this study aimed to determine the probability of the occurrence of a near-
miss crossing situation event because it was not possible to analyse the actual
collision event, due to the difficulty of finding complete data on collision accidents.
Because the AIS data obtained is also limited, therefore two time-intervals are
drawn from the whole data set to capture the condition of Sunda Strait during 2018-
2019, when TSS has not been set.
The analysis here is conducted for two time intervals as previously mentioned in
Table 4-1 and each time interval is going through analyses with three different bases.
Figure 4.1 Acceptance criteria for individual risk based (UK HSE, no date).
58
Firstly, an analysis for the crossing zone basis is presented in Table 4-2. The results
show that in the Time Interval 1, only Crossing Zone 6 and Crossing Zone 7 are
found as acceptable, meanwhile Crossing Zone 1 – 4 are unacceptable and the
Crossing Zone 5 is located in the ALARP level. At the Time Interval 2, CZ 4
remained in the unacceptable level, while the CZ 1 changed to ALARP level.
The next one is still using time-based for the course basis and the result is given in
Table 4-3. As shown in the table, the crossing frequency of Course 1 experienced a
significant increase in Time Interval 3, which is after the implementation of TSS.
This is due to no near-miss found in Time Interval 1 and Time Interval 2 at all.
While at Time Interval 3, five events are found. The application of TSS as well as
associated routeing measures such as the precautionary area in the intersection area
between Course 3 and Course 4 which is traversed by roro/passenger ferries with
Course 1 and Course 2 which is traversed by cargo ships which are not accompanied
by additional rules regarding the priority of ships that can pass through
precautionary area causes an increase in near-miss situation. In addition, the good
impact of implementing TSS is felt by the decrease in the frequency of near-miss
Table 4-2 The result of time-based probability assessment for

the crossing zone basis.
Crossing zone The num. The num. of Annual

Probability
no. of trips near miss frequency
Time Interval 1
CZ 1 172 2 2.33E-02 1.57E-03
CZ 2 108 1 1.85E-02 1.25E-03
CZ 3 205 4 3.90E-02 2.64E-03
CZ 4 430 13 6.05E-02 4.09E-03
CZ 5 368 2 1.09E-02 7.36E-04
CZ 6 63 0 0.00E+00 0.00E+00
CZ 7 91 0 0.00E+00 0.00E+00
Time Interval 2
CZ 1 260 1 7.69E-03 2.90E-04
CZ 2 126 0 0.00E+00 0.00E+00
CZ 3 305 0 0.00E+00 0.00E+00
CZ 4 791 16 4.05E-02 1.52E-03
CZ 5 500 5 2.00E-02 7.54E-04
CZ 6 59 0 0.00E+00 0.00E+00
CZ 7 80 0 0.00E+00 0.00E+00
59
the course basis.
The num. The num. of Annual

Vessel course Probability
of trips near miss frequency
Time Interval 1
Course 1 17 0 0.00E+00 0.00E+00
Course 2 21 2 9.52E-02 6.45E-03
Course 3 328 17 5.18E-02 3.51E-03
Course 4 341 5 1.47E-02 9.92E-04
Course 5 92 7 7.61E-02 5.15E-03
Course 6 91 8 8.79E-02 5.95E-03
Course 7 32 3 9.38E-02 6.35E-03
Course 8 36 2 5.56E-02 3.76E-03
Course 9 144 0 0.00E+00 0.00E+00
Course 10 119 0 0.00E+00 0.00E+00
Time Interval 2
Course 1 27 0 0.00E+00 0.00E+00
Course 2 22 0 0.00E+00 0.00E+00
Course 3 538 15 2.79E-02 1.05E-03
Course 4 542 7 1.29E-02 4.87E-04
Course 5 92 10 1.09E-01 4.10E-03
Course 6 99 11 1.11E-01 4.19E-03
Course 7 32 1 3.13E-02 1.18E-03
Course 8 31 0 0.00E+00 0.00E+00
Course 9 98 0 0.00E+00 0.00E+00
Course 10 69 0 0.00E+00 0.00E+00
crossing situations experienced in Course 7 and Course 8 which were initially at

unacceptable level, then decreased to ALARP for Course 7 and acceptable for
Course 8.
Lastly, the result of vessel type basis shown that the annual frequency of container
and fishing vessels become unacceptable after TSS because these types of vessels
tend to be undiscoverable during the Time Interval 1 and Time Interval 2. While
the annual frequency for roro/passenger ferries decreased to acceptable level.
4.5 Root Cause Analysis
The risk assessment process begins with a frequency assessment and then continues
with a consequence assessment. Consequence assessment can be done with
qualitative or quantitative methods, depending on the degree of interest of the study.
60
the vessel type basis.

Type of vessel Probability
Time Interval 1
Bulk carrier 78 8 1.03E-01 6.94E-03
Chemical tanker 24 5 2.08E-01 1.41E-02
Container ship 10 0 0.00E+00 0.00E+00
Crude oil tanker 3 1 3.33E-01 2.26E-02
Fishing vessel 1 0 0.00E+00 0.00E+00
Gas tanker 15 1 6.67E-02 4.51E-03
General cargo 43 2 4.65E-02 3.15E-03
Oil products tanker 110 4 3.64E-02 2.46E-03
Roro / Passenger ferry 808 23 2.85E-02 1.93E-03
Pleasure vessel 0 0 0.00E+00 0.00E+00
Support vessel 104 0 0.00E+00 0.00E+00
Other vessel 25 0 0.00E+00 0.00E+00
Time Interval 2
Bulk carrier 78 7 8.97E-02 3.38E-03
Gas tanker 26 1 3.85E-02 1.45E-03
General cargo 47 5 1.06E-01 4.01E-03
Other vessel 24 2 8.33E-02 3.14E-03
However, many studies put forward quantitative assessment in order to get a vivid
picture of the level of risk being faced. In this study, the quantitative method has
been applied to calculate the frequency of near-miss situations caused by high
marine traffic in Sunda Strait. Furthermore, the consequence assessment is carried
out qualitatively using the bow tie method with the aim of communicating risks
involving cause-consequence relationships in a system on a strategic level (de
Ruijter and Guldenmund, 2016).
61
4.5.1 Bow tie analysis
In general, two approaches are known on how the bow tie analysis is being utilized:
quantitatively and qualitatively. The concept of both approaches is same as it is
using the bow tie diagram (BTD) to model the relationship of causes and
consequences which is connected by a top event. Figure 4.2 displays the basic
concept of BTD. While having the same concept, the quantitative bow tie analysis
combines two well-known risk analysis methods, fault tree analysis (FTA) and
event tree analysis (ETA), that also including safety barriers (Ehlers et al., 2017).
The FTA is used in the left side of the BTD to calculate the probability of a critical
event or a top event to happen, and the ETA is utilising in the right side to find the
likelihood of each consequence to exist given all the safety barriers. On the other
hand, the qualitative approach is rather a simple cause–effect relationship analysis
which has all possible causes in the left side that could lead to a critical event. This
approach is considered to be useful “when the situation does not warrant the
complexity of a full fault tree analysis or when the focus is more on ensuring that
there is a barrier or control for each failure pathway” (ISO, 2019). These causes are
joined with OR gates as in the FTA. While the right side provides all consequences
that can be caused directly by the top event (de Ruijter and Guldenmund, 2016).
Safety barriers or safety measures are divided into two categories, the prevention
and mitigation measures. The prevention or avoid measures located on left side
which means measures that can prevent the top event to happen due to the
associated cause. While the mitigation measure on the right side shows what action
that could be done so that the critical event happened would not result to the
associated consequence.
This study attempted to conduct a consequence assessment not to find out what
consequences that could be generated, but rather as a way to know all causes that
could lead to a top event. Furthermore, a safety barrier model is constructed to
prevent the hazard to be developed and resulted on a collision. The universal picture
of the BTD of collision accident in Sunda Strait could be described by Figure 4.3.
To be able to develop and conclude the bow tie diagram in the figure below, several
62
Figure 4.2 Basic concept of bow tie diagram (Ehlers et al., 2017).
accident reports that happened in Sunda Strait are analysed and modelled by using
the Variation Tree Analysis (VTA). The VTA is used to see the sequence of an
event and get the idea on which sequence the deviation or variation is formed. The
variation factors can be defined when the sequence departs from the safe procedure
(Tamura and Shinoda, 2009). An example of the report that was analysed is the
collision involving a gas tanker Norgas Cathinka and roro ferries Bahuga Jaya that
happened on September 26th, 2012, in Sunda Strait and the VTA is shown in Figure
4.4. All threats and departed events are extracted from the accident reports through
VTA and then categorised into four groups of threats. Firstly, external factors that
do not directly occur on board the ship but can give an influence on the decisions
taken by the mariners. The second one is management related factors, which related
to the safety management system applied by the ship operators or the ship owners.
Next is the factors contributed by unskilled seafarers. Lack of knowledge, skills,

and physical barriers are included in this group. Lastly, technical problems that have
direct effect to the operation of the ship. All categories of threats are concluded
based on real collision accidents report which happened before the implementation
of TSS, so it is more general and involving more factors outside TSS itself.
63
Figure 4.3 Bow tie analysis for crossing collision in Sunda Strait.
4.5.2 Safety Barrier for Collision Accidents

Since bow tie analysis requires a more in-depth analysis to determine the barriers
that can avoid a threat to turn into a top event, a safety barriers model for collision
accidents is proposed in this study. According to the PSA (2013), the safety barrier
is defined as “systems of technical, operational and organizational elements, which
are intended individually or collectively to reduce the possibility for a specific error,
hazard or accident to occur, or which limit its harm/disadvantages”. The
development of safety barrier initially is based on the study by Reason (1990) that
discovered a Swiss Cheese Model to model a human error so as it can result on an
accident. In the maritime sector, several models of safety barriers have been
published, for example a model for safety barriers in a ship capsizing accident (Wu
et al., 2017) and for offshore drilling blowout (Xue et al., 2013). In this occasion, a
safety barriers model for vessel collision is explored as shown in Figure 4.5, while
the causal categories are explained in detail to better understand the root causes that
can make the barriers failed.
Safety barriers for vessel collision are divided into seven groups. Four of them are
used to prevent threats that have been explained in the bow tie analysis, while the
rest are applied when the initial event or the top event, which in this case is the near-
miss situation, has been initiated. These three barriers could prevent each stage of
collision before it develops further to be a real collision accident, and namely risk
64
Figure 4.4 VTA of collision between Norgas Cathinka and Bahuga Jaya.
65
Figure 4.5 Safety barriers for collision accidents.
of collision barrier, close quarter barrier, and immediate danger barrier. The root
causes of each barrier are developed based on the departed event of VTA that have
been previously discussed. During the extraction of the root causes, any speculative
factors are excluded from this method. Those seven barriers are elaborated in Table
4-5 and Table 4-6.
The effectivity of TSS could contribute to the safety barrier and could prevent the
causal categories to happen, especially at EB 1: no local regulation about the
priority to pass the precautionary area (PA), EB 5: poor traffic control system, SB
1: poor knowledge about the ship, saling area, COLREGs, and TSS and SB 4: poor
passage planning.
66
Table 4-5 Details of safety barriers (Part 1)
Code Barriers No Causal categories

EB External barrier EB1 No local regulation about the priority to
pass the precautionary area (PA)
EB2 Inconsistency of local regulations
EB3 Incomplete weather information
EB4 Poor environmental condition (shallow,
rapid current, heavy traffic, narrow)
EB5 Poor traffic control system (equipment,
standard operating procedure, piloting
procedure)
MB Management barrier TB1 Poor company safety management system
TB2 No periodical safety training
TB3 Incompliance with the SOLAS
SB Seafarers barrier SB1 Poor knowledge about the ship, sailing area,
COLREGs, and TSS
SB2 Poor navigational skill
SB3 Poor bridge resource management
SB4 Poor passage planning
SB5 Poor physical condition
TB Technical barrier PB1 Machineries failure
PB2 Poor condition of the safety equipment
PB3 Poor condition of the navigational
equipment
PB4 Undone previous maintenance task
PB5 Faulty design of ship
RB Risk of collision barrier RB1 Improper lookout (visual, equipment,
underkeel clearance)
RB2 Neglect to watch (talk with others)
RB3 Poor assumption (based on scanty
information)
RB4 Failed to recognize the risk of collision
RB5 Lack of situational awareness
RB6 Failed to maintain a safe speed & safe
distance
CB Close-quarter barrier CB1 Failed to judge the role of the ship (stand-on
/ give-way)
CB2 Poor judgment about the surrounding
condition
CB3 Failed in ship-to-ship communication (low
volume, language barrier)
CB4 Failed in ship-to-port communication
CB5 Failed to have a sight of other vessel
67
Table 4-6 Details of safety barriers (Part 2)
Code Barriers No Causal categories

IB Immediate danger IB1 Give-way vessel not showing collision
barrier avoidance action
IB2 Failed to take collision avoidance action
(stand-on vessel late to maneuver)
IB3 Failed to take collision avoidance action
(incompliance with COLREGs)
68
Chapter 5
TSS as a Risk Control Option
The process of determining risk control options (RCOs) is made to reduce the level
of risk by analysing all factors that contribute to the high level of risk. According
to IMO MSC-MEPC.2/Circ.12/Rev.2 in Appendix 6, there are three categories of
risk control measures (RCMs): category A (preventive and mitigating), category B
(engineering, inherent, and procedural), and category C (diverse, redundant, passive,
independent, dependent, involved human factors, critical human factors, auditable
or not auditable, quantitative or qualitative, established or novel, developed or non-
developed). The RCMs are grouped into limited number of RCO to be adapted in
the practical regulation.
The following paragraphs explain about the risk control option in the form of ships’
routeing measures that is implemented in Sunda Strait as an effort to reduce the risk
of collision in all crossing zones. As the result of risk assessment shows the number
of annual frequencies in Crossing Zone 1, 2, 3, and 4 are in the unacceptable level
during Time Interval 1, therefore reduction of risk of collision can be performed by
reducing the annual frequency of the near-miss.
5.1 Ships’ Routeing Measures
According to the book of ships’ routeing published by IMO, the purpose of routeing
measures is initially for safety purpose, but now the purpose is expanded to the
protection of marine environments and ecologies (IMO, 2010). As safety also have
69
Figure 5.1 (a) Traffic separation scheme separated by separation zones in 4 and
separation line in 3, (b) inshore traffic zone adjacent with other routeing systems,
(c) precautionary area with a recommended traffic flow (IMO, 2010).
a corelation to the safety of people’s live, the ships’ routeing is also contained in
the SOLAS convention Chapter V about Safety Navigation in Regulation 10. There
are 16 terms that are used to describing the routeing systems as defined in the book
of ships’ routeing, and at least four of them are implemented in Sunda Strait, those
are depicted in Figure 5.1 and explained as follows:
1. Traffic separation scheme (TSS), which is defined as a routeing system that is

meant to separate the opposing traffic by establishing different traffic lanes.
2. Separation zone or line mean a zone or line that located between two traffic
lanes which has ships in opposite directions, or it can also separate a traffic lane
from the adjacent sea area, else it can be used as a mean for separating a certain
type of vessels which proceeding in a same direction. Either separation zone or
70
separation line can be used to separate the opposing traffic, but a separation
zone is preferred as long as there is no narrow or restricted channel.
3. Inshore traffic zone (ITZ) is a specific area situated in the traffic separation
scheme boundary with the adjacent coast and shall not be used when vessels
can use TSS safely. However, smaller vessels which have length less than 20
meters, sailing vessels and vessels engaged in fishing may use the inshore
traffic zone. In addition, vessels sailing from or to the port, offshore structure,
or other places that located near the ITZ can also sail in this area as stated in
the COLREGs Rule 10 (d).
4. Lastly, the precautionary area which is translated as a location where ships that
sailing through this area must pay a thorough attention and caution. A traffic
flow recommendation could be given to this area to increase the safety of
navigation.
In the relation with proposed routeing systems, before IMO can adopt a new or
amend the existing one, it should consider about the preparedness of the proposing
states by looking at two aspects: supporting aids to navigation (AtoN) that can help
mariners while navigating in this area, adequate hydrographic surveys. An aspect
is added while assessing a proposed TSS which is the compliance with the design
criteria as regulated by IMO.
In the following paragraphs, two examples of TSS are explained as a comparison

to routeing systems implemented in Sunda Strait. The first one is TSS Dover Strait
that is noted as the first ever TSS adopted by IMO. The other one is TSS Strait of
Malacca and Singapore (SOMS) which located close to TSS Sunda Strait.
5.2 TSS in other countries
5.2.1 TSS Dover Strait

This TSS is located between two states, England and France and was initially
implemented as a voluntary domestic regulation in 1967 before it was adopted by
IMO and became mandatory to all vessels passing this TSS. This scheme as
depicted by Figure 5.2 was proposed originally due to the high number of collisions
71
Figure 5.2 TSS Dover Strait (Sea-Seek, 2012)
that caused massive losses (Dare and Lewison, 1980). The situation in Dover Strait
is similar with the Sunda Strait, as it also has a heavy crossing traffic in the north
part of the strait that later became the background of setting a precautionary area
with traffic flow recommendation around this area, as previously there was a
conflict between Rule 10 with Rule 15 and Rule 17 of COLREGs. A seminar was
held to discuss about practical problems faced while navigating in the Dover Strait
and it resulted on a termination of TSS and replacing it with a precautionary area to
increase the degree of freedom of crossing vessels, so the risk of collision is not
developed any further.
The development of TSS Dover Strait is followed by the establishment of ship

reporting systems. At first, a voluntary ship reporting called MAREP is used to
assist all vessels under 300 GT to increase safety of navigation in the English
Channel and Dover Strait and minimise the risk of pollution off the coasts of the
United Kingdom and France. Later, a new mandatory ship reporting system named
CALDOVREP which is applied to ships above 300 GT to replace the MAREP as
72
outlined in the SN/Circ.167, annex, page 4, amending paragraph 7 of the
Recommendations on navigation through the English Channel and Dover Strait.
5.2.2 TSS Strait of Malacca and Singapore

The Strait of Malacca and Singapore (SOMS) is one of the busiest channels in the
world because this strait connects the Indian Ocean to the Pacific Ocean. This strait
is a host to approximately 80,000 vessels annually as well as 80% of the global
trades (Qu and Meng, 2012). Figure 5.3 shows the number of ships that give report
to the Klang VTS which is located in Malaysia. Three states (i.e., Singapore,
Malaysia, and Indonesia) are involved in the maritime cooperation regime in the
SOMS through a forum called Tripartite Technical Expert Group on the Safety of
Navigation (TTEG) in 1975. This forum has produced two results, the first one is a
traffic separation scheme in 1981 and the second one is a mandatory ship reporting
system (STRAITREP) in 1998 (Abdul Rahman, 2014), which depicted in Figure
5.4.
Figure 5.3 Number of ships reporting to Klang VTS by type

(Jabatan Laut Malaysia, 2021).
73
Figure 5.4 TSS Strait of Malacca and Singapore (Zhen, 2016).
An urgency to set a TSS in the SOMS is due to a severe traffic as this strait is the
shortest path that could possibly links Gulf of Middle East to the East Asia. Most
countries in East Asia are reckoned markets for oil and gas produced by countries
in Middle Eastern countries, which make the transportation here is vital to their
energy supply chain. When the traffic in SOMS has reached their maximum
capacity, the similar case could also happen to Sunda Strait and Lombok strait due
to the strategic position of these two straits, as shown in Figure 5.5. More vessels
will divert their course to go through Sunda Strait to maintain their time schedule.
This opportunity is caught by Indonesia and hence two TSSs were proposed to IMO
in 2019.
5.3 TSS in Sunda Strait Indonesia
During the 6th meeting of the Navigations, Communications, and Search and
Rescue (NCSR) sub-committee at IMO headquarter in January 2019, two proposals
74
Figure 5.5 Choke-points, sea lines and mega-hub
around Indian and Pacific Ocean (Rimmer and Lee, 2007).
proposed by the Government of Indonesia (GoI) were approved. The proposals

contain TSS and associated routeing measures that have objectives to reduce the
risk of collision accidents and grounding incident. After they were approved, the
proposals were adopted by the Maritime Safety Committee (MSC) in their 101 st
meeting in June 2019. Afterwards, TSS along with associated routeing measures
were implemented at July 1st, 2020 for both Sunda Strait and Lombok Strait (IMO,
2019).
75
While TSS is one of the routeing measures recognised by IMO, in Sunda Strait itself
there are more than one routeing measures implemented, which are traffic
separation scheme with separation zone and separation line, inshore traffic zone,
and precautionary area. The objectives of setting those routeing measures in Sunda
Strait as written in the proposal are to: 1) decrease the number of head-on situations
by separating opposing traffic, 2) decrease the risk of ship grounding by keeping
ships away from the coral reef atolls of Terumbu Koliot, 3) minimise the risk of
ship collision by recommending one precautionary area, and 4) make sure that ships
following the course that free of any dangers due to the coral reef atolls of Terumbu
Koliot by mean of aids to navigation (AToN) (IMO, 2018a).
5.3.1 Properties of TSS Sunda Strait

The properties and position of complete routeing measures in Sunda Strait are
shown in Figure 5.6 and the coordinates of each point in that figure is presented in
Table 5-1. The following paragraph gives a brief description of each feature.
1. Northern TSS has 2.67 nm in wide and 2.6 nm in length is separated by a

separation zone with 0.3 nm wide. It is situated exactly in the north side of the
precautionary area.
2. A precautionary area which has recommendations for traffic flow direction is
located between the northern TSS and the southern TSS. The northern entrance
has 2.67 nm wide, while the south entrance has 1.2 nm wide.
3. Southern TSS has 1.2 nm wide in the northern entrance, which is located next
to the precautionary area. Meanwhile, the southern entrance has a wider
entrance with 4 nm. This TSS does not have a separation zone, rather separation
line as the passage is narrower compared to the north bound TSS.
4. Inshore traffic zones which bounded by TSS and the east coast of the Sumatera
Island (Tg. Kanggalan and Tg. Tua).
5. Lastly, SUNDAREP or the ship reporting system (SRS) in TSS Sunda Strait.
This feature is not drawn in the map because this is a supporting feature that
shall be made by Indonesian flagged ships sailing in or crossing TSS. This
reporting system is also recommended to be performed by foreign flagged ships.
76
The SRS is meant to manage the traffic, to ensure the safety of navigation, as
well as to protect the marine environment.
As previously depicted in Figure 3.2, the traffic pattern in Sunda Strait previously
is irregular because there is no regulation that set out a specific course for every
vessel. The only available course is the IASL 1, which is an imaginary line. In fact,
in the common practice, the seafarers will make a passage plan that is not always
following the IASL 1. Oftentimes, a ship will sail in a diagonal direction as
represented by Course 7 and Course 8. After the implementation of TSS, a ship
navigating in Sunda Strait has to choose whether she is going to follow TSS or
sailing through the eastern part of Sunda Strait, or Course 5 and Course 6.
To enforce TSS and all supporting routeing measure in the national level, the GoI
has enacted a Ministerial Decree of the Ministry of Transportation No. 13 of 2020
which has been implemented since July 1st, 2020. This regulation sets out all
requirements and regulation that shall be obeyed by all Indonesian vessels and
foreign vessels that sail through TSS Sunda Strait.
Figure 5.6 Properties of routeing measures in Sunda Strait.
77
5.3.1.1 Traffic comparison before and after TSS
The implementation of TSS has greatly affected the traffic pattern in Sunda Strait.
Before applying TSS, the traffic is quite random, but still some main channels can
be observed as displayed in Figure 5.7. This figure has a heatmap of the traffic
during March 1st until April 19th, 2019. There are at least six courses that can be
clearly seen in Sunda Strait. The first one is having a north-south direction in the
western part of Sunda Strait, where this line is the axis of IASL 1 and where TSS is
located now. The second path is a through passage which has the same direction as
the first but is close to the Port of Merak. The third course is the crossing channel
which is used for roro ferries with the route from Merak to Bakauheni and vice
versa, having a high density. The next one is a diagonal passage that cuts directly
Sunda Strait from the southwest to the northeast. In addition, a quite clear traffic
from the southeast is also spotted, which is heading towards the northwest, but in
the middle of the roro ferries routes, its direction changed to the north. The last one
is a crossing course located on the south side of Sunda Strait which leads to or
originates from the south side of the island of Java. Due to the very dense and
diverse traffic in Sunda Strait as a result of passing cargo ships from Australia or
Europe heading to East Asia or vice versa, which is caused by the presence of IASL
1, the risk of ship collision is high. Several accidents have occurred in this area that
Table 5-1 Coordinates of TSS Sunda Strait.
No. Coordinates No. Coordinates

1 05°48.30' S, 105°50.35' E 9 05°53.34' S, 105°48.06' E
2 05°48.89' S, 105°51.31' E 10 05°53.97' S, 105°49.09' E
3 05°49.06' S, 105°51.58' E 11 05°55.03' S, 105°48.43' E
4 05°49.66' S, 105°52.54' E 12 05°54.41' S, 105°47.39' E
5 05°51.94' S, 105°51.13' E 13 05°56.38' S, 105°45.51 E
6 05°51.34' S, 105°50.16' E 14 05°57.76' S, 105°47.32' E
7 05°51.17' S, 105°49.89' E 15 05°53.65' S, 105°48.56' E
8 05°50.57' S, 105°48.92' E 16 05°57.04' S, 105°46.46' E
78
causing considerable losses. Consequently, it is necessary to regulate the traffic to
reduce this risk.
The ratification of TSS Sunda Strait by the Government of Indonesia since July 1 st,
2020, made the traffic through Sunda Strait more organized, especially in areas on
the west side of Sunda Strait where the opposing traffics are separated. Even so, we
can still see an accumulation of traffic that occurs around the precautionary area, as
shown in Figure 5.8. This heatmap shows traffic conditions in Sunda Strait during
April 6th – July 24th, 2021, a year after TSS was firstly implemented. From this
figure, the separation zone and line clearly separate the traffic on the IASL 1 line.
However, the effectiveness of these routing systems still needs to be analysed to
find out whether there is a true reduction in the risk of ship collision after the
implementation of TSS, which will be discussed in the next chapter.
Port of Bakauheni
!
Port of Suralaya
Port of Merak
Port of Tanjung Gerem
Figure 5.7 Heatmap before the implementation of TSS.
79
Port of Bakauheni
Port of Suralaya
Port of Merak
Port of Tanjung Gerem
Figure 5.8 Heatmap after the implementation of TSS.
5.4 Assessing the Effectivity of TSS as an RCO
Implementation of the routeing measures, specifically TSS and the precautionary

area. As mandated by IMO Instruments Implementation Code (III Code) through
IMO Resolution A.1070(28) that coastal states have obligations to “periodically
evaluate its performance in respect of exercising its rights and meeting its
obligations under the applicable international instruments”. Hence, the probability
and the annual frequency of near-miss crossing situations are analysed considering
the number of near-miss that happened over time to provide a better picture of the
traffic condition after the implementation of TSS Sunda Strait.
In relation to this study, the trip-based frequency analysis that have been done for
the crossing zone, course, and vessel type bases produced a frequency level that
have been elaborated on in this study in Chapter 4.2. The results of risk assessment
can be written as:
80
1. The calculation of annual frequency during Time Interval 1 produced a
result that showing a severe risk of collision in the Crossing Zone 1, 2, 3,
and 4. While in Time Interval 2, unacceptable level only found in Crossing
Zone 4 and ALARP level is found in Crossing Zone 1 and 5. The result
indicated that a high-density traffic that is found in those crossing zones
could lead to a potential of collision accident.
2. Next, the result of course basis presented that almost all courses are in
unacceptable level, except Course 1, 9, and 10 that are acceptable and
Course 4 which is on ALARP level. Meanwhile, Time Interval 2 resulted
the same level for all, except Course 2 that now became acceptable.
3. Lastly, the analysis using the vessel type basis. The risk level between Time
Interval 1 and Time Interval 2 has similar results. Only container, fishing
vessel, pleaser vessel, support vessel and other vessel are on acceptable level
at Time Interval 1. The changes were experienced by roro/passenger ferries
that changed from unacceptable to ALARP level in Time Interval 2 and for
the other vessel that was previously acceptable in Time Interval 1 became
unacceptable in Time Interval 2.
In order to prove the effectivity of the routeing systems proposed for Sunda Strait
as depicted by Figure 5.6, the risk assessment step that already performed in
Chapter 4 is performed again. The purpose is to see a change in the level of the risk,
or at least a decrease in the value of annual frequency of near-miss situations. A
time interval is added to represent the condition after the implementation of TSS
and named as Time Interval 3. The time-based analysis is still be used in this risk
assessment, because 7 days are also taken for Time Interval 3. Hence, the same
equations are utilised to get the probability and the annual frequency of the near-
miss situations in Sunda Strait. The information of the new time interval is exhibited
in Table 5-2. Although the vessel counts and trip counts are not similar between
before and after the implementation of TSS, but the value of trips per vessel per
hour of all time intervals are similar. This number assumes that the traffic density
in each period is almost the same. Hence, a comparison for annual frequency can
be made for these three intervals.
81
Table 5-2 Information about new time intervals.
Data set no. Time Interval 1 Time Interval 2 Time Interval 3

Start date and time October 8th, 2018 April 13th, 2019 July 1st, 2021
01:16:17 GMT 00:00:18 GMT 00:00:31
End date and time October 14th, 2018 April 19th, 2019 July 7th, 2021
04:44:48 GMT 09:35:30 23:59:15 GMT
Time duration 106 hours 150 hours 168 hours
Vessel counts 295 301 544
Trip counts 1221 1550 2668
Trip/(vessel.hour) 0.0390 0.0343 0.0292
Following the same process as Chapter 4.4 about the calculation of probabilistic
risk assessment, a new result is produced for Time Interval 3 for all bases as shown
in Table 5-3, Table 5-4, and Table 5-5.
Table 5-3 gives the result of crossing zone basis that indicates no crossing zone falls
under the unacceptable level after the implementation of TSS (Time Interval 3).
Reductions in the risk level are experienced by Crossing Zone 2, 3, and 4 during
Time Interval 3 compared to the Time Interval 1. In addition, the annual frequency
in CZ 1 and 2 reduced to ALARP, and CZ 3 became acceptable. These explanations
shows that TSS is effective to reduce the risk of collision of all crossing zones.
The result of the course basis before and after TSS is displayed in Table 5-4. It
shows that the annual frequency in Course 1 rises drastically when the traffic
follows TSS. This situation was expected as the traffic must follow the direction of
TSS as well as the precautionary area where the traffic is likely to concentrate in
this area. The traffic in Course 1 definitely passes the precautionary area and may
be facing traffic concentration that could lead to a crossing situation with vessels
sailing in Course 3 or Course 4. Another result shows although the crossing
situation in Course 3 and Course 4 are a lot higher than in Course 1, but the annual
frequency is less vulnerable because the rules set in those courses are less strict
82
the crossing zone basis before and after TSS implementation.
Crossing The num. The num. of Annual

Probability
zone no. of trips near miss frequency
Time Interval 1
CZ 1 172 2 2.33E-02 1.57E-03
CZ 2 108 1 1.85E-02 1.25E-03
CZ 3 205 4 3.90E-02 2.64E-03
CZ 4 430 13 6.05E-02 4.09E-03
CZ 5 368 2 1.09E-02 7.36E-04
CZ 6 63 0 0.00E+00 0.00E+00
CZ 7 91 0 0.00E+00 0.00E+00
Time Interval 2
CZ 1 260 1 7.69E-03 2.90E-04
CZ 2 126 0 0.00E+00 0.00E+00
CZ 3 305 0 0.00E+00 0.00E+00
CZ 4 791 16 4.05E-02 1.52E-03
CZ 5 500 5 2.00E-02 7.54E-04
CZ 6 59 0 0.00E+00 0.00E+00
CZ 7 80 0 0.00E+00 0.00E+00
Time Interval 3
CZ 1 834 8 1.92E-02 3.75E-04
CZ 2 866 3 6.93E-03 1.35E-04
CZ 3 794 0 0.00E+00 0.00E+00
CZ 4 1038 19 3.66E-02 7.15E-04
CZ 5 716 12 3.35E-02 6.55E-04
CZ 6 74 0 0.00E+00 0.00E+00
CZ 7 164 0 0.00E+00 0.00E+00
compared to courses bounded by TSS, so vessels can navigate more freely, but still
in the frame if COLREGs, to avoid the near-miss crossing situation. Lastly, the
annual frequency in Course 7 became ALARP and Course 8 became unacceptable
which mean they are reduced significantly. The reason is because the vessel that
previously navigating in Course 7 and 8 before TSS, are then moving to follow the
course of TSS. This reduction can be deducted as the indicator of the effectivity of
TSS.
Finally, Table 5-5 revealed the calculation result of the vessel type basis. The annual
frequency of bulk carrier, chemical tanker, and roro/passenger ferries reduced to
ALARP which brings us a supportive statement that TSS is effective. Meanwhile,
83
the course basis before and after TSS implementation.

Vessel course Probability
Time Interval 1
Course 1 17 0 0.00E+00 0.00E+00
Course 2 21 2 9.52E-02 6.45E-03
Course 3 328 17 5.18E-02 3.51E-03
Course 4 341 5 1.47E-02 9.92E-04
Course 5 92 7 7.61E-02 5.15E-03
Course 6 91 8 8.79E-02 5.95E-03
Course 7 32 3 9.38E-02 6.35E-03
Course 8 36 2 5.56E-02 3.76E-03
Course 9 144 0 0.00E+00 0.00E+00
Course 10 119 0 0.00E+00 0.00E+00
Time Interval 2
Course 1 27 0 0.00E+00 0.00E+00
Course 2 22 0 0.00E+00 0.00E+00
Course 3 538 15 2.79E-02 1.05E-03
Course 4 542 7 1.29E-02 4.87E-04
Course 5 92 10 1.09E-01 4.10E-03
Course 6 99 11 1.11E-01 4.19E-03
Course 7 32 1 3.13E-02 1.18E-03
Course 8 31 0 0.00E+00 0.00E+00
Course 9 98 0 0.00E+00 0.00E+00
Course 10 69 0 0.00E+00 0.00E+00
Time Interval 3
Course 1 88 5 5.68E-02 1.11E-03
Course 2 89 6 6.74E-02 1.32E-03
Course 3 730 21 2.88E-02 5.62E-04
Course 4 732 20 2.73E-02 5.34E-04
Course 5 272 16 5.88E-02 1.15E-03
Course 6 267 15 5.62E-02 1.10E-03
Course 7 70 1 1.43E-02 2.79E-04
Course 8 53 0 0.00E+00 0.00E+00
Course 9 259 0 0.00E+00 0.00E+00
Course 10 108 0 0.00E+00 0.00E+00
the annual frequency of container, fishing vessel, general cargo, and oil products
tanker remain in unacceptable level. The reason of it could be in conjunction with
the lack of data that makes them undiscoverable during Time Interval 1 and Time
Interval 2. Besides that, the analysis of the vessel type basis is performed for all
courses, which also producing result for all courses. Hence, if result for a focused
course or crossing zone is desired, then an analysis must be done separately.
84
the vessel type basis before and after TSS implementation.

Type of vessel Probability
Time Interval 1
Bulk carrier 78 8 1.03E-01 6.94E-03
Gas tanker 15 1 6.67E-02 4.51E-03
General cargo 43 2 4.65E-02 3.15E-03
Other vessel 25 0 0.00E+00 0.00E+00
Time Interval 2
Bulk carrier 78 7 8.97E-02 3.38E-03
Gas tanker 26 1 3.85E-02 1.45E-03
General cargo 47 5 1.06E-01 4.01E-03
Other vessel 24 2 8.33E-02 3.14E-03
Time Interval 3
Bulk carrier 127 5 3.94E-02 7.69E-04
Container ship 24 2 8.33E-02 1.63E-03
Crude oil tanker 36 0 0.00E+00 0.00E+00
Fishing vessel 29 2 6.90E-02 1.35E-03
Gas tanker 87 3 3.45E-02 6.74E-04
General cargo 113 11 9.73E-02 1.90E-03
Support vessel 319 3 9.40E-03 1.84E-04
Other vessel 27 1 3.70E-02 7.24E-04
After observing all results of the time-based analysis for all bases, a conclusion that
could be drawn is a declining trend of annual frequency in almost all crossing zone,
85
which lead to a statement that TSS is effective. However, an extra attention shall be
paid to Course 1 and Course 2, where precautionary area is located and Course 5
and 6 where the traffic is severe. As the risk level in those courses are prone to the
change and could easily become unacceptable.
The fact that TSS itself is a RCO that is expected to reduce the number of collisions
in an area with a dense traffic, based on results generated by the risk assessment
process above, three additional RCOs are proposed which considered to be able to
further reduce the risk level to be acceptable or ALARP.
5.4.1 Enforcement of COLREGs

According to the Ministerial Decree of Ministry of Transportation no. 130 of 2020
stated that all actions taken in TSS Sunda Strait must be in compliance with
COLREGs. COLREGs Rule 10(c) said that any vessels engaged in TSS shall avoid
crossing traffic lanes, and COLREGs Rule 15 stated that the vessel encountering
others in her starboard shall give way and avoid crossing ahead other vessel, if
possible. These rules shall be obeyed by all mariners navigating in Sunda Strait,
especially those engaged in TSS. However, several crossing situations observed
even after the implementation of TSS shows that they were incompliance with
COLREGs Rule 15 and Rule 10. Hence, all parties involved in this condition must
enforce the application of COLREGs in all situations and the government shall take
any actions to warn the contravening vessels, so their operation will not give more
threat to the marine traffic of Sunda Strait.
5.4.2 Priority in Precautionary Area

The first and second results of the trip-based analysis stated that there is an increase
in the annual frequency for crossing zones and courses passing the precautionary
area. This RCO still has a corelation with the first RCO, because the action taken
in the PA shall follow COLREGs either Rule 10 or 15. But, cargo vessels that pass
through TSS often take the stand on role. It would be better if an additional
regulation is set to prioritise the vessel engaged in TSS to pass first, if a crossing
situation happens with the roro/passenger ferries. So, the risk of near-miss that
could lead to a collision accident can be reduced.
86
5.4.3 Routeing systems in Eastern Side of Sunda Strait
As explained in the first result of the frequency assessment that the crossing
frequency in the CZ 5 doubled after the implementation of TSS and located in the
unacceptable level. To be able to regulate the traffic in this location, new routeing
systems can be implemented so a reduction of the annual frequency levels are
expected.
5.5 Cost Benefit Assessment
The purpose of conducting a cost benefit assessment (CBA) is to furnish a brief

comparison between the potential benefits and costs related to the implementation
of RCOs and the potential losses when the RCOs are not implemented. Together
with that, this assessment is also be used to prioritise which RCO that can have
more resources than the others (Viertola and Storgard, 2013). The cost-benefit
methodology concerns about the evaluation of public projects which the outcomes
are evaluated based on the public interest. In addition, the cost and benefit are
assessed in terms of social utility gains and losses rather than in a form of cashflows
(Nas, 2016).
A study suggested that an accident could generate to an immediate result, either an

injury or death that will have a direct or indirect cost to the company (DeMarco,
Frederick and Thomas, 1997). The direct cost is related to some amount of money
that has to be paid to cover the fatalities or any injuries caused by the accident, such
as insurance premium, lawsuits, medical expenses, etc. While the indirect costs
involve the side effect the accident that usually are not easily seen, for example
disruptions in the workplace, productivity loss, attorney fee, etc. The worst part is
that the amount of indirect cost actually could be twice or even more than the direct
cost (Morrison, 2014). Despite of the definition of direct and indirect cost that could
potentially charge a lot to the company, a single fatality can never be quantified in
terms of money and the process of monetising the death could give a severe impact
to the company. As for the unquantified costs and benefits can be discussed
qualitatively and relative importance can be drawn based on expert judgements.
87
However, the FSA is an attempt to evaluate a current or new policy implemented
to the public, which in this case is all vessels passing Sunda Strait, that should have
a clear result, so the decision makers have a better understanding to which RCO
that should be taken. A method such as “willingness to pay” is often used to
calculate the amount of money that a party put to reduce the risk. Another approach
is by comparing the economic benefit of an RCO with the cost to establish and
maintain the RCO (Polyzois, 2017). Two methods to calculate the cost-
effectiveness are known in the process of CBA in order to make the result of FSA
to more consistent and easily comparable and understandable, i.e., GCAF (Gross
Cost of Averting a Fatality) and NCAF (Net Cost of Averting a Fatality). The
GCAF indicates the ratio of additional cost of the RCO to reduce the risk in terms
of fatalities averted, while the NCAF puts the economic benefits of the RCO into
account. To be categorised as a cost-effective, the value of either CAF needs to be
under 3 million US dollar, based on IMO guidelines for FSA. The equation used to
calculate those two indices are:
(15)
(16)
In the case of TSS in Sunda Strait as an RCO, only the cost of infrastructure is
considered. Based on a speech by Head of the Navigation District Class I Dumai,
Raymond Ivan H. Santuri on the International Workshop on The Designation of
Traffic Separation Scheme (TSS) and Its Associated Routeing Measures in Sunda
and Lombok Straits at Hotel Aryaduta Jakarta dated June 30th, 2019, the proposing
state shall have established infrastructures that can support the daily operational of
TSS. The explanation of each is briefly discussed in the following paragraphs.
5.5.1 Cost due to the Implementation of RCO

1. Vessel Traffic System (VTS)
According to the SOLAS Chapter 5 Regulation 12 and IMO SN Circ. 296 a
vessel traffic system must be established by proposing state to monitor the
88
traffic condition in the area of TSS, which consists of Workstation, Radar,
Automatic Identification System (AIS), Radio VHF Communication, CCTV,
Meteorologic Sensor, Binoculars, Electronic charts, VTS data system,
Communication link system, and Electrical system.
2. Coastal Radio Station
As TSS Sunda Strait is also supported by a mandatory ship reporting system
(SRS) that is called as SUNDAREP, it needs a coastal radio station to help the
communication of ship that want to report their position to the shore authority
and that include Setting the GMDSS (Global Maritime Distress and Safety
System), VHF Transceiver System, MF/HF Transceiver System, Operator
Station, Operator Supporting System, Back Up Power System, and Installation
& mobilisation service.
3. Aids to Navigation (AtoN)
As mandated by SOLAS Chapter 5 Regulation 13 to build AtoN to help the
vessels that navigating within TSS area. The equipment that shall be
established in relation to the AtoN are Lighthouse, Beacons, Lantern House,
Lantern Mast With Secured Battery Box (Large), Revolving Lantern, Solar
Module, VRLA Absorbent Glass Mate (AGM), Lightning rod, Surge Arrester
Module, Power Supply & AtoN Controller Panel (Standard), Diesel generator,
and Tetrahedral Aluminum Radar Reflector.
4. Electronic Chart
This device is used to show the chart of Sunda Strait onboard the ship.
In this research, two cases for the cost assessment are studied. As Sunda Strait
already has VTS, the establishment of VTS shall not be included into the calculation.
In Case 1, the set-up of a coastal radio station is not needed, but the establishment
of aton is required. While in Case 2 a coastal radio station is needed, as well as the
aids to navigation. Lastly, both cases do not consider the electronic chart into
account as the chart can be made virtual and all vessels must already have it
installed onboard. All the consideration mentioned is displayed in the Table 5-6. In
this table, the total cost of each case is also calculated and shown.
89
Table 5-6 Comparison of two cases of the cost assessment.
No. Infrastructure cost Cost Case 1 Case 2

1 Vessel Traffic System (VTS) 4,863,157 - -
2 Coastal radio station 469,614 - +
3 Aids to navigation (AtoN) 453,470 + +
4 Electronic chart 10,807 - -
Total cost (in USD) 5,797,048 453,470 923,084
5.5.2 Benefits resulted by the Implementation of RCO

The benefits outlined here are the risk reduction that can be obtained by
implementing TSS. Two scenarios are considered in this assessment. The first one,
the risk reduction is looked at the whole crossing zones. While the second scenario
only consider the crossing zone that directly affected by TSS.
There are two parameters that can be used to calculate the ∆𝑅 of the implementation
of TSS: reduction of annual frequency and reduction rate. The reduction is defined
by the difference between the annual frequency and Time Interval 1 is used as the
basis value for the annual frequency before TSS is implemented and compare it
with Time Interval 3 after the setting of TSS. Equations for reduction and reduction
rate are outlined as follow:
∆𝑅 = 𝐹𝐶𝑍𝑖 𝑜𝑓 𝑇𝑖𝑚𝑒 𝐼𝑛𝑡𝑒𝑟𝑣𝑎𝑙 1 − 𝐹𝐶𝑍𝑖 𝑜𝑓 𝑇𝑖𝑚𝑒 𝐼𝑛𝑡𝑒𝑟𝑣𝑎𝑙 3 (17)
∆𝑅
𝑅𝑒𝑑𝑢𝑐𝑡𝑖𝑜𝑛 𝑅𝑎𝑡𝑒(%) = 𝐹 (18)
𝐶𝑍𝑖 𝑜𝑓 𝑇𝑖𝑚𝑒 𝐼𝑛𝑡𝑒𝑟𝑣𝑎𝑙 1
Following those equations, the value for Scenario 1 and 2 are calculated and it is
found in Table 5-7 that the number of reduction rate of the Scenario 1 is 82% and
in Scenario 2 is 91%. It means that TSS is more effective when the analysis is only
considering the area where TSS is applied. Although a high reduction rate is also
experienced by all CZi after the implementation of TSS. The risk reduction is finally
chosen as the parameter representing the ∆𝑅 in this study.
90
Table 5-7 Comparison of two scenarios of the benefit assessment.
No. Duration All CZi (Scenario 1) CZ 1, CZ 2, and CZ

3 directly related to
TSS (Scenario 2)
1 Time Interval 1 1.03E-02 5.46E-03
2 Time Interval 3 1.88E-03 5.10E-04
Reduction 8.41E-03 4.95E-03
Reduction rate (%) 82% 91%
compared to TI 1
5.5.3 Result of Cost Benefit Assessment

After knowing that there are two cases in the cost assessment as well as in the
reduction risk which is considered as the benefit of the implementation of TSS, the
calculation of GCAF as in Equation 15 is proceeded. This parameter is chosen
because it does not involve the calculation of economic benefits of TSS, as it could
lead to a bias and subjectivity from the assessor. The result is contained in Table
5-8 and from this table it can be deduced that the lowest value of GCAF is available
for Case 1 for crossing zones that have relation with TSS (i.e., CZ 1, 2, and 3) with
the value of 500,191 USD. The result of this analysis gives a rough view to the
implementation of TSS that will be most cost effective for the Case 1 when the
infrastructure to be built is only aids to navigation and dedicated for the area of TSS.
If this case is looked from the perspective of all traffic in Sunda Strait, it would be
less cost-effective because the number of GCAF increased by 54,697 USD.
Table 5-8 Results of cost benefit assessment.
RCO Case Expected cost GCAF GCAF

(△C in USD) (△R for All CZi) (△R for CZ 1,2,3)
Case 1 453,470 554,888 500,191

Case 2 923,084 1,129,532 1,018,190
91
5.6 Recommendation for Decision Maker
The study of safety assessment here is used to understand that the implementation
of TSS Sunda Strait can reduce the risk of near-miss crossing situation. However,
the risk cannot be reduced significantly, at least in a short period as TSS has just
implemented in Sunda Strait for 2 years. The result of this assessment shows that
Course 1, Course 2, Course 5 and Course 6 are located on the unacceptable level,
which shall be reduced by implementing the RCOs. The risk control options must
be employed sooner, and the traffic must be recorded afterwards. As mandated by
IMO Instruments Implementation Code (III Code) through IMO Resolution
A.1070(28) that coastal states must regularly evaluate the performance of the newly
implemented policy, which in this case is TSS. So, traffic monitoring must be
conducted consistently, and any situations must be recorded by the VTS. The fact
that the VTS Merak does not have an ability to get the record of AIS data in Sunda
strait as per now, an enhancement to the VTS Merak must be carried out.
92
Chapter 6
Conclusion and Future Research
Efforts to reduce the number of marine accidents has been conducted recently and
are starting to show positive results. A dense marine traffic can produce a hazard
and result in ship collisions. Therefore, the marine traffic must be managed
properly. One measure that can be applied is the implementation of ship routing
systems. Traffic separation scheme is a routing system that can separate opposing
traffics. Sunda Strait as one of the busiest waters in Indonesia where IASL 1 is
located has just implemented this scheme starting from July 1st, 2020. By
implementing TSS, the risk of head-on collision is expected to experience a
decrease. Even so, the risk of collisions can still be found along with the discovery
of several crossing zones in Sunda Strait. This research has an aim to evaluate the
collision risk in the area of TSS Sunda Strait by means of safety assessment. It has
four objectives and this chapter summarize the findings that can be drawn with
regard to the research objective.
6.1 Research Findings
1. Develop a framework for a vessel near-miss crossing situation database in

Sunda Strait by utilizing the available Automatic Identification System (AIS)
data.
This research has addressed the problem of vessel collision due to the high traffic
density in Sunda Strait. A framework of safety assessment is proposed that consists
of four steps: hazard identification, risk assessment, determination of RCO, and
recommendation for decision makers. The framework to deal with AIS data
93
handling is also proposed. There are few academic studies that discussed about TSS
as a measure to reduce the risk of collision, so this study has contributed to the field
of safety assessment, especially for TSS.
2. Evaluate the risk of vessel collision in Sunda Strait before and after TSS is
implemented.
A risk assessment is performed as a mean to prove that TSS has successfully

reduced the risk of collision. A probability assessment is carried out twice, before
and after TSS is implemented by using time-based and trip-based approaches. The
assessment uses three perspectives: crossing zone, course, and vessel type basis,
which never been used to assess the risk of collision. So, this research presents a
novelty on the risk assessment approach. Interesting results are given by this study
about how TSS could affect the annual frequency of the near-miss crossing situation.
Although a reduction is observed in almost of the scenario, but an increase is also
found in several course. Hence, a continuous attention shall be paid to the traffic
condition to avoid the increment of annual frequency any further. The levels of
annual frequency are spanned from acceptable, ALARP, to unacceptable level.
3. Identify the root causes of the vessel collision by mean of Bow tie analysis.
The risk level of the collision accident is already known, the root causes of the
vessel collision is identified in Chapter 4.5. Some novelties are offered in this
chapter, such as the use of bow tie analysis to see the cause-consequence
relationship that could produce a vessel collision. In addition, the root causes can
be known by exploring the safety barriers. Seven orders of safety barriers for vessel
collision are proposed here which has never been elaborated by another research.
4. Propose risk control options to decrease the number of vessel collision risk
based on the identified root causes.
Three risk control options to the existing TSS Sunda Strait are offered to further
reduce the level of risk in TSS Sunda Strait. Alongside, recommendations are also
given to continuously monitor the marine traffic in Sunda Strait. A renewal and
94
enhancement to the VTS Merak shall be carried to, so they can provide a better
service to maintain the safety of navigation in Sunda Strait.
5. Assess the cost effectivity of TSS as a risk control option.
The effectivity of TSS in terms of its cost is assessed. This assessment found that
the implementation of TSS is effective if it is seen from the perspective of all
crossing zones in TSS (i.e., CZ 1, 2, and 3). The value of GCAF for this case is
500,191 USD and this value is the lowest compared to another case.
6. Give recommendations to the Government of Indonesia to improve the

safety of navigation in Sunda Strait.
As the final step of the safety assessment framework, recommendations are given
to the Government of Indonesia which are based on the result of the assessment. It
contains the obligation to monitor the traffic condition in Sunda Strait and followed
up by recording all near-miss situation by local VTS. Other recommendations is to
set up further RCOs that is based on the results of risk assessment.
6.2 Future Research
The improvement could be performed during the process of AIS data handling.
Instead of excluding the missing AIS data, a data imputation method will provide
another picture of the marine traffic of Sunda Strait. Another data set could also be
added to the analysis to enrich the result and make it more precise. The future
research about the AIS data handling is an interesting topic, as AIS can be
categorised as big data and a lot of information is provided by AIS. So, if the
information inside the AIS message is maximised, a thorough study about risk
assessment could be performed. Beside the AIS data handling, a validation of the
trips extracted by IWRAP could also be improved by utilising a machine learning
by classifying the course based on their course and coordinates. So, the course of
the vessels can be predicted although the collected AIS data is not complete.
95
The next one is related to the risk assessment, the implementation of RCOs is
expected to further reduce the frequency of the near-miss crossing situation. A
simulation-based analysis could be developed to prove that the RCO could give a
positive impact to the marine traffic in Sunda Strait. In line with the previous topic,
the consequence analysis could also be further studied. Therefore, a solid risk
assessment could be mapped in a risk matrix and risk acceptance criteria can be
drawn for crossing zone, course, as well as vessel type basis.
Finally, the implementation of TSS which initially is proposed to reduce the

collision accident. But as the research about autonomous vessels is now being
emerged, it is not impossible that TSS could be a trend in near future because it is
considered to ease up the operation of the autonomous vessels.
96
Reference
Abdul Rahman, N. (2014) ‘Multilateralism in the Straits of Malacca and Singapore’,
Marine Policy, 44, pp. 232–238. doi:10.1016/j.marpol.2013.08.030.
Baldauf, M. and Wiersma, E. (1998) ‘Risk Recognition and Collision Avoidance

by VTS Operators’, 7th IFAC Symposium on Analysis, Design and Evaluation of
Man Machine Systems (MMS’98), Kyoto, Japan, 16-18 September 1998, 31(26), pp.
239–244. doi:10.1016/S1474-6670(17)40099-1.
Bošnjak, R., Šimunović, L. and Kavran, Z. (2012) ‘Automatic Identification System

in Maritime Traffic and Error Analysis’, Transactions on Maritime Science, 1(2),
pp. 77–84. doi:10.7225/toms.v01.n02.002.
Cem Kuzu, A., Akyuz, E. and Arslan, O. (2019) ‘Application of Fuzzy Fault Tree
Analysis (FFTA) to maritime industry: A risk analysing of ship mooring operation’,
Ocean Engineering, 179, pp. 128–134. doi:10.1016/j.oceaneng.2019.03.029.
Chin, H.C. and Debnath, A.K. (2009) ‘Modeling perceived collision risk in port
water navigation’, Safety Science, 47(10), pp. 1410–1416.
doi:10.1016/j.ssci.2009.04.004.
Coello, J. et al. (2015) ‘An AIS-based approach to calculate atmospheric emissions

from the UK fishing fleet’, Atmospheric Environment, 114, pp. 1–7.
doi:10.1016/j.atmosenv.2015.05.011.
Coldwell, T.G. (1983) ‘Marine Traffic Behaviour in Restricted Waters’, Journal of

Navigation. 2009/11/23 edn, 36(3), pp. 430–444.
doi:10.1017/S0373463300039783.
Čorić, M. et al. (2021) ‘Quantitative Ship Collision Frequency Estimation Models:

A Review’, Journal of Marine Science and Engineering, 9(5).
doi:10.3390/jmse9050533.
CSIS (2020) Maritime Safety in Indonesia. Centre for Strategies and International
Studies. Available at:
https://www.csis.or.id/uploads/attachments/post/2020/01/17/CSIS_Maritime_Safe
ty_Report.pdf (Accessed: 8 December 2021).
97
CTSB (2021) Marine Transportation Occurrences in 2020. Statistical Summary.
Transportation Safety Board of Canada. Available at: https://www.bst-
tsb.gc.ca/eng/stats/marine/2020/ssem-ssmo-2020.pdf (Accessed: 24 November
2021).
Cutlip, K. (2017) ‘AIS for Safety and Tracking: A Brief History’, Global Fishing
Watch, 31 March. Available at: https://globalfishingwatch.org/data/ais-for-safety-
and-tracking-a-brief-history/ (Accessed: 10 December 2021).
D. Liu and G. Shi (2020) ‘Ship Collision Risk Assessment Based on Collision
Detection Algorithm’, IEEE Access, 8, pp. 161969–161980.
doi:10.1109/ACCESS.2020.3013957.
Dare, S. and Lewison, G. (1980) ‘The Recent Casualty Record in the Dover Strait’,
Journal of Navigation, 33, pp. 192–205. doi:10.1017/S0373463300035207.
DeMarco, G., Frederick, R. and Thomas, C. (1997) The Economic Impat of

Accidents on the Marine Industry. National Information Service - Springfield (Tech.
No. PB98-111420).
Dinariyana, A.A.B. (2020) ‘Traffic Separation Scheme di Selat Sunda dan Selat
Lombok’. Pembekalan Tentang Traffic Separation Scheme (TSS) Pelayaran di
Selat Sunda dan Selat Lombok, 24 August.
Ehlers, U.C. et al. (2017) ‘Assessing the safety effects of cooperative intelligent
transport systems: A bowtie analysis approach’, Accident Analysis & Prevention,
99, pp. 125–141. doi:10.1016/j.aap.2016.11.014.
European Maritime Safety Agency (2019) Annual Overview of Marine Casualties

and Incidents 2019.
Fadhil, M., Handani, D.W. and Dinariyana, A.A.B. (2020) ‘Risk Analysis on
Modified Offloading System of LNG FSRU’, in IOP Conf. Series: Earth and
Environmental Science. Maritime Safety International Conference, IOP Publishing.
doi:doi:10.1088/1755-1315/557/1/012020.
Friis-Hansen, P. (2008) ‘Basic Modelling Principles for Prediction of Collision and

Grounding Frequencies’, Technical University of Denmark, p. 59.
Fujii, Y. and Shiobara, R. (1971) ‘The Analysis of Traffic Accidents’, Journal of

Navigation, 24(4), pp. 534–543. doi:10.1017/S0373463300022372.
Fukuto, J. and Imazu, H. (2013) ‘New collision alarm algorithm using obstacle zone
by target (OZT)’, IFAC Proceedings Volumes, 46(33), pp. 91–96.
doi:10.3182/20130918-4-JP-3022.00044.
Goodwin, E.M. (1975) ‘A Statistical Study of Ship Domains’, Journal of

Navigation, 28(3), pp. 328–344. doi:10.1017/S0373463300041230.
98
Görçün, Ö.F. and Burak, S.Z. (2015) ‘Formal Safety Assessment for Ship Traffic
in the Istanbul Straits’, 11th International Strategic Management Conference, 207,
pp. 252–261. doi:10.1016/j.sbspro.2015.10.094.
Hammer, W. (1972) Handbook of System and Product Safety. Prentice Hall.
Heidarzadeh, M. et al. (2020) ‘Field Survey of Tsunami Heights and Runups

Following the 22 December 2018 Anak Krakatau Volcano Tsunami, Indonesia’,
Pure and Applied Geophysics, 177(10), pp. 4577–4595. doi:10.1007/s00024-020-
02587-w.
Hilgert, H. (1983) ‘Defining the Close-Quarters Situation at Sea’, Journal of

Navigation. 2009/11/23 edn, 36(3), pp. 454–461.
doi:10.1017/S0373463300039801.
ICS (2020) Shipping and world trade: driving prosperity. International Chamber of
Shipping. Available at: https://www.ics-shipping.org/shipping-fact/shipping-and-
world-trade-driving-prosperity/ (Accessed: 23 November 2021).
IMO (1998) Adoption of New and Amended Performance Standards, Resolution

MSC.74(69).
IMO (ed.) (2010) Ships’ routeing. Tenth edition. London: International Maritime
Organization.
IMO (2018a) ‘Establishment of a new traffic separation scheme and associated

routeing measures in the Sunda Strait, Indonesia’.
IMO (2018b) Revised Guidelines for Formal Safety Assessment (FSA) for Use in
the IMO Rule-Making Process, MSC-MEPC.2/Circ.12/Rev.2.
IMO (2019) ‘Sub-Committee on Navigation, Communications and Search and

Rescue (NCSR), 6th session, 16-25 January 2019’, 25 January. Available at:
https://www.imo.org/en/MediaCentre/MeetingSummaries/Pages/NCSR-6th-
session.aspx (Accessed: 7 December 2021).
IMO (2021) ‘Maritime Safety Committee (MSC)’, 24 November. Available at:

https://www.imo.org/en/MediaCentre/MeetingSummaries/Pages/MSC-
Default.aspx (Accessed: 24 November 2021).
IMO (no date) ‘AIS transponders’. Available at:

http://www.imo.org/en/OurWork/Safety/Navigation/Pages/AIS.aspx (Accessed:
27 November 2019).
Indonesian Navy (2013) ‘Indonesia Jawa Bagian Barat’.
IPC (no date) ‘The anchor position docking area (Anchorage Area)’. Indonesia Port
Corporations.
99
van Iperen, E. (2015) ‘Classifying Ship Encounters to Monitor Traffic Safety on
the North Sea from AIS Data’, TransNav, the International Journal on Marine
Navigation and Safety of Sea Transportation, 9(1), pp. 51–58.
doi:10.12716/1001.09.01.06.
ISO (2019) Risk management — Risk assessment techniques, IEC 31010:2019.
Jabatan Laut Malaysia (2021) ‘Statistik Pelayaran’, November. Available at:

https://www.marine.gov.my/jlm/halaman/pelayaran-6 (Accessed: 14 January
2022).
JICA (2003) The Study for Development of Greater Jakarta Metropolitan Ports.
Summary Report, p. 30. Available at:
https://openjicareport.jica.go.jp/pdf/11744232_02.pdf (Accessed: 7 December
2021).
JTSB (2021) Statistics of Marine Accident. Statistics. Japan Transport Safety Board.
Available at: https://www.mlit.go.jp/jtsb/statistics_mar.html (Accessed: 24
November 2021).
Kaplan, S. and Garrick, B.J. (1981) On The Quantitative Definition of Risk.

Available at: https://doi.org/10.1111/j.1539-6924.1981.tb01350.x.
Kontovas, C. and Psaraftis, H. (2009) ‘Formal Safety Assessment: A Critical

Review’, Marine Technology, 46, pp. 45–59. doi:10.5957/mtsn.2009.46.1.45.
Kristiansen, S. (2013) Maritime transportation: safety management and risk

analysis. Routledge.
Last, P., Hering-Bertram, M. and Linsen, L. (2015) ‘How automatic identification

system (AIS) antenna setup affects AIS signal quality’, Ocean Engineering, 100,
pp. 83–89. doi:10.1016/j.oceaneng.2015.03.017.
Le Tixerant, M. et al. (2018) ‘How can Automatic Identification System (AIS) data
be used for maritime spatial planning?’, Maritime Spatial Planning, Ecosystem
Approach and Supporting Information Systems (MapSIS 2017), 166, pp. 18–30.
doi:10.1016/j.ocecoaman.2018.05.005.
Learnmarine (2018) ‘Radar Plotting or... Do we really understand what ARPA

does?’, 17 August. Available at: https://www.learnmarine.com/blog/Radar-
Plotting-or...-Do-we-really-understand-what-ARPA-does (Accessed: 18 December
2021).
Lenart, A. (2017) ‘Approach Parameters in Marine Navigation – Graphical

Interpretations’, TransNav, the International Journal on Marine Navigation and
Safety of Sea Transportation, 11(3), pp. 521–529. doi:10.12716/1001.11.03.19.
100
Lin, B. and Huang, C.-H. (2006) ‘Comparison between arpa radar and AIS
characteristics for vessel traffic services’, Journal of Marine Science and
Technology, 14, pp. 182–189. doi:10.51400/2709-6998.2072.
Liu, B. et al. (2021) ‘Tracing illegal oil discharges from vessels using SAR and AIS
in Bohai Sea of China’, Ocean & Coastal Management, 211, p. 105783.
doi:10.1016/j.ocecoaman.2021.105783.
Lu, L. et al. (2020) ‘Critical aspects for collision induced oil spill response and
recovery system in ice conditions: A model-based analysis’, Journal of Loss
Prevention in the Process Industries, 66, p. 104198. doi:10.1016/j.jlp.2020.104198.
Macduff, T (1974) ‘The Probability of Vessel Collisions’, Ocean Industry, 9(9), pp.
144–148.
MAIB (2021) Marine Accident Recommendations and Statistics 2020. Annual

Report. Marine Accident Investigation Branch, p. 53. Available at:
https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attac
hment_data/file/992017/MAIBAnnualReport2020.pdf (Accessed: 24 November
2021).
Maritime Administrator (2021) Annual Report on Marine Safety Investigations

2020. Republic of the Marshall Islands Maritime Administrator, p. 9. Available at:
https://safety4sea.com/wp-content/uploads/2021/04/RMI-Annual-Report-on-
Marine-Safety-Investigations-2020-2021_04.pdf (Accessed: 24 November 2021).
Melchior, B. and Baumert, K. (2014) Indonesia: Archipelagic and other Maritime

Claims and Boundaries. 141. Office of Ocean and Polar Affairs Bureau of Oceans
and International Environmental and Scientific Affairs U.S. Department of State.
Available at: https://www.state.gov/wp-content/uploads/2020/02/LIS-141.pdf
(Accessed: 7 December 2021).
Ministry of Transportation (2014) Setting the Size of Crossing Roro Ferries in

Merak-Bakauheni (in Bahasa Indonesia: Pengaturan Ukuran Kapal Angkutan
Penyeberangan di Lintas Merak-Bakauheni), Ministerial Regulation of the
Ministry of Transportation No. 88 of 2014. Available at:
http://jdih.dephub.go.id/produk_hukum/view/VUUwZ09EZ2dWRUZJVlU0Z01q
QXhOQT09 (Accessed: 7 December 2021).
Ministry of Transportation (2019) The obligation to install and activate AIS devices
in Indonesian waters (in Bahasa Indonesia: Pemasangan dan Pengaktifan Sistem
Identifikasi Otomatis bagi Kapal yang Berlayar di Wilayah Perairan Indonesia),
Ministerial Decree of the Ministry of Transportation No. 7 of 2019.
Miyake, R., Fukuto, J. and Hasegawa, K. (2015) ‘Analyses of the Collision

Avoidance Behaviours Based on AIS Data’, The Journal of Japan Institute of
Navigation, 133, pp. 66–74. doi:10.9749/jin.133.66.
101
Morrison, K.W. (2014) ‘The ROI of Safety’, Safety and Health Magazine, 23 May.
Available at: https://www.safetyandhealthmagazine.com/articles/10414-the-roi-of-
safety (Accessed: 24 December 2021).
Mujeeb-Ahmed, M.P. and Paik, J.K. (2021) ‘Quantitative collision risk assessment
of a fixed-type offshore platform with an offshore supply vessel’, Structures, 29,
pp. 2139–2161. doi:10.1016/j.istruc.2020.06.026.
Nas, T.F. (2016) Cost-benefit analysis: Theory and application. Lexington Books.
Nieh, C. et al. (2014) ‘Evaluation of imputation methods for microbial surface

water quality studies’, Environmental Science: Processes & Impacts, 16(5), pp.
1145–1153. doi:10.1039/C3EM00721A.
NMEA (2011) ‘Automatic Identification Systems’. National Marine Electronics

Association. Available at:
https://www.nmea.org/Assets/nmea%20collision%20avoidance%20through%20ai
s.pdf (Accessed: 10 December 2021).
Norway delegates (no date) FORMAL SAFETY ASSESSMENT: Decision

parameters including risk acceptance criteria Submitted by Norway, MSC 72/16.
NTSC (2012) Tubrukan antara MT. Gloria Sentosa dengan Kapal Jukung Irpansya.
Investigasi Kecelakaan Kapal Laut KNKT-11-03-03-03. Komite Nasional
Keselamatan Transportasi Republik Indonesia (National Transportation Safety
Committee).
NTSC (2013) Investigation Into the Collision Between Indonesia Registered Ro-ro
Passenger Ferry MV “Bahuga Jaya” and Singapore Registered MT “Norgas
Cathinka” at Sunda Strait (4 Nautical miles east of Rimau Balak Island), Indonesia.
KNKT-12-09-03-03.
NTSC (2016) Grounding of HANJIN AQUA (IMO No. 9632480) At Terumbu Koliot,
Sunda Strait, Merak. Marine casualty investigation report KNKT.15.12.09.03.
NTSC (2018) Senggolan Mutiara Persada I dengan Port Link III. Final
KNKT.17.07.20.03. Komite Nasional Keselamatan Transportasi Republik
Indonesia (National Transportation Safety Committee).
NTSC (2020) Senggolan Srikandi Indonesia dengan Maestro Diamond dan Angel
No. 1. Final KNKT.17.08.24.03. Komite Nasional Keselamatan Transportasi
Republik Indonesia (National Transportation Safety Committee).
Park, J.W. (2014) A Study on the Estimation of a Safe Speed for Prevention of Ship’s
Collision in VTS Area. Mokpo Maritime University.
Polyzois, A. (2017) A Cost-benefit Analysis of the Risk Control Options towards

Mitigating Maritime Hazards Leading to Injuries and Fatalities in Tanker Shipping
102
Companies: a Critical and Thought Provoking Case study. World Maritime
University. Available at:
https://commons.wmu.se/cgi/viewcontent.cgi?article=1547&context=all_dissertati
ons (Accessed: 24 December 2021).
PSA (2013) Principles for Barrier Management in the Petroleum Industry.

Stavanger, Norway: Petroleum Safety Authority.
Qu, X. and Meng, Q. (2012) ‘The economic importance of the Straits of Malacca
and Singapore: An extreme-scenario analysis’, Select Papers from the 19th
International Symposium on Transportation and Traffic Theory, 48(1), pp. 258–
265. doi:10.1016/j.tre.2011.08.005.
Rahim, G.P. and Sunaryo (2019) ‘Container Ship Accident Analysis due to
Container Stacked on Deck as an Attempt to Improve Maritime Logistic System’,
in E3S Web of Conferences. IC-AMME 2018, EDP Sciences.
doi:https://doi.org/10.1051/e3sconf/201913001010.
Reason, J. (1990) Human error. New York, NY, US: Cambridge University Press
(Human error.), pp. xv, 302.
Rimmer, P. and Lee, P. (2007) ‘Repercussions of impeding shipping in the Malacca

and Singapore Straits’, Journal of International Logistics and Trade, 5.
doi:10.24006/jilt.2007.5.1.001.
de Ruijter, A. and Guldenmund, F. (2016) ‘The bowtie method: A review’, Safety

Science, 88, pp. 211–218. doi:10.1016/j.ssci.2016.03.001.
Sakar, C. et al. (2021) ‘RISK ANALYSIS OF GROUNDING ACCIDENTS BY

MAPPING A FAULT TREE INTO A BAYESIAN NETWORK’, Applied Ocean
Research, 113, p. 102764. doi:10.1016/j.apor.2021.102764.
Schofield, C. and Arsana, I.M.A. (2009) ‘Closing the loop: Indonesia’s revised
archipelagic baselines system’, Australian Journal of Maritime and Ocean Affairs,
1(2), pp. 57–62.
Sea-Seek (2012) ‘Pas de Calais - Dover Straight’. Sea-seek. Available at:

https://www.sea-seek.com/en/Pas-de-Calais-Dover-Straight (Accessed: 17
December 2021).
Silveira, P. et al. (2021) ‘A multicriteria outranking approach for ship collision risk
assessment’, Reliability Engineering & System Safety, 214, p. 107789.
doi:10.1016/j.ress.2021.107789.
Squire, D. (2003) ‘The hazards of navigating the Dover Strait (Pas-de-Calais) traffic
separation scheme’, The journal of navigation, 56(2), pp. 195–210.
103
Tamura, Y. and Shinoda, T. (2009) ‘Safety Assessment with Marine Casualty Data
on Collision Accidents between Fishing Vessels and Cargo Vessels’, in.
UK HSE (no date) Guidance on ALARP Decisions in COMAH,

SPC/Permissioning/37. Available at:
https://www.hse.gov.uk/foi/internalops/hid_circs/permissioning/spc_perm_37/.
Viertola, J. and Storgard, J. (2013) Overview on the Cost-Effectiveness of Maritime

Safety Policy Instruments. A 66. University of Turku.
Wang, J. and Foinikis, P. (2001) ‘Formal safety assessment of containerships’,

Marine Policy, 25(2), pp. 143–157. doi:10.1016/S0308-597X(01)00005-7.
Wu, B. et al. (2017) ‘An Evidential Reasoning-Based CREAM to Human

Reliability Analysis in Maritime Accident Process: An Evidential Reasoning-Based
CREAM to Human Reliability Analysis’, Risk Analysis, 37. doi:10.1111/risa.12757.
Xue, L. et al. (2013) ‘A safety barrier-based accident model for offshore drilling
blowouts’, Journal of Loss Prevention in the Process Industries, 26(1), pp. 164–
171. doi:10.1016/j.jlp.2012.10.008.
Yanti, R., Aviolita, S. and Marsetio (2020) ‘Indonesia also feels the urgency to set
the TSS in its water, because Indonesia has contributed to the great number of
accidents worldwide.’, International Journal of Law and Political Sciences, 14(12).
Zhang, M. et al. (2021) ‘A method for the direct assessment of ship collision
damage and flooding risk in real conditions’, Ocean Engineering, 237, p. 109605.
doi:10.1016/j.oceaneng.2021.109605.
Zhen, S. (2016) ‘Regulation on Navigation of Foreign Vessels – Singapore’s

Practice’. 4th Asia Pacific Ocean Law Institutions Alliance Conference, 22 June.
Available at: https://cil.nus.edu.sg/wp-content/uploads/2016/07/APOLIA-Sun-
Zhen.pdf (Accessed: 17 December 2021).
104
Appendix A
Near-miss Situation Database
Based on the DCPA calculation explained in Chapter 2.2.2 that was conducted on
the IWRAP, numerous near-miss situations were detected and recorded on the
Figure A1 – Figure A4. Information on each data set is contained in the following
tables.
Data set 1 (Before implementation of TSS)

Start date and time October 8th-14th, 2018
01:16:17 GMT – 04:44:48 GMT
End date and time April 13th – 19th, 2019
00:00:18 GMT – 09:35:30
Vessel counts 504
Trip counts 2771
Near-miss counts 22 events (October 8th-14th, 2018)
22 events (April 13th – 19th, 2019)
Data set 2 (After implementation of TSS)

Start date and time July 1st, 2021
00:00:31
End date and time July 7th, 2021
23:59:15 GMT
Vessel counts 544
Trip counts 2668
Near-miss counts 42 events
105
Figure A1 Near-miss crossing situations for Data set 1 (Part 1).
106
Taking action
Crossing Heading of Heading of COG of COG of Avg. speed Avg. speed Role of Role of
Action taken based on
num. Vessel A Vessel B Vessel A Vessel B of Vessel A of Vessel B Vessel A Vessel B
COLREG?
1 34 30 10 8.74 Give-way Stand-on Vessel A give-way and changed the heading to the left to avoid crossing ahead Vessel B Yes
2 257 290 249 8.51 15.97 Give-way Stand-on Vessel B give-way and changed the heading to the left to avoid crossing ahead Vessel A No
3 291 216 287 212 9.4 20.1 Give-way Stand-on Vessel A give-way and changed the heading to the right to avoid crossing ahead Vessel B Yes
5 35 39 39 4.31 12.71 Give-way Stand-on Vessel A give-way and changed the heading to the right to avoid crossing ahead Vessel B Yes
6 64 305 60 9.58 5.9 Stand-on Give-way Vessel B give-way and changed the heading to the right to avoid crossing ahead Vessel A Yes
7 259 39 269 40 10.33 10.68 Stand-on Give-way Vessel A give-way and changed the heading to the left to avoid crossing ahead Vessel B No
8 259 17 258 7 8.12 9.68 Stand-on Give-way Vessel A give-way and changed the heading to the left to avoid crossing ahead Vessel B No
13 81 213 100 203 4.81 9.81 Stand-on Give-way Vessel B give-way and changed the heading to the right to avoid crossing ahead Vessel A Yes
16 308 225 291 222 9.86 15.89 Give-way Stand-on Vessel B give-way and changed the heading to the left to avoid crossing ahead Vessel A No
19 299 271 69 9.15 11.69 Stand-on Give-way Vessel B give-way and changed the heading to the left to avoid crossing ahead Vessel A Yes
20 93 213 110 212 5.8 21.6 Stand-on Give-way No significant change is observed No
107
26 202 293 202 9.31 10.09 Give-way Stand-on No significant change is observed No
27 272 42 277 43 12.45 7.32 Stand-on Give-way Vessel A give-way and changed the heading to the right to avoid crossing ahead Vessel B No
28 283 202 273 206 15.68 10.34 Give-way Stand-on Vessel B give-way and changed the heading to the right to avoid crossing ahead Vessel A No
33 32 267 32 7.28 14.62 Stand-on Give-way Vessel A give-way and changed the heading to the right to avoid crossing ahead Vessel B No
38 207 118 211 7.42 4.9 Stand-on Give-way No significant change is observed No
39

188 189 285 7.75 5.39 Give-way Stand-on Vessel A give-way and changed the heading to the right to avoid crossing ahead Vessel B Yes
40 63 283 63 12.03 14.41 Stand-on Give-way No significant change is observed No
41 273 37 261 37 8.72 8.41 Stand-on Give-way Vessel A give-way and changed the heading to the right to avoid crossing ahead Vessel B No
44 211 272 210 8.71 11.57 Give-way Stand-on Vessel B give-way and changed the heading to the left to avoid crossing ahead Vessel A No
Heading Heading
Crossing Time start MMSI of MMSI of Course of Course of
Date Loc. Vessel A Vessel B Type of Vessel A Type of Vessel B of Vessel of Vessel
num. (GMT) Vessel A Vessel B Vessel A Vessel B
A B
1 01-Jul-21 02:06:47 CZ 5 KMP MENGGALA HELMY REJEKI 525024157 525019467 Roro / Passenger ferry Oil products tanker 4 5
2 01-Jul-21 07:52:44 CZ 1 NUSA PUTERA PRINCESS A 525005357 538008572 Roro / Passenger ferry Bulk carrier 4 2 21
3 01-Jul-21 09:30:52 CZ 4 NEOMI GOLDEN DESTINY 525200924 525100148 Roro / Passenger ferry General cargo 3 6 279
4 01-Jul-21 09:51:26 CZ 5 PORT LINK III ERASAMUDRA5 525005177 525600670 Roro / Passenger ferry Fishing vessel 4 5
5 01-Jul-21 09:51:26 CZ 5 PORT LINK III GOLDEN DESTINY 525005177 525100148 Roro / Passenger ferry General cargo 4 6
6 01-Jul-21 10:04:19 CZ 2 PORTLINK V PALUH TABUAN/PERTAMINA 1019 525005232 525008013 Roro / Passenger ferry Oil products tanker 3 1 40
7 01-Jul-21 10:37:34 CZ 5 WIRA BERLIAN SUN TRITON 525300391 374316000 Roro / Passenger ferry Oil products tanker 4 6 109 235
8 01-Jul-21 10:49:02 CZ 4 NUSA PUTERA MARITIME VOYAGER 525005357 564145000 Roro / Passenger ferry General cargo 3 5 27
9 01-Jul-21 12:37:44 CZ 5 MITRA NUSANTARA KANAKA 05 525002063 525020111 Roro / Passenger ferry Support vessel 4 6 211
10 01-Jul-21 13:51:20 CZ 4 AMADEA CEMPAKA 525201093 525019002 Roro / Passenger ferry Other vessel 4 6 130 267
11 01-Jul-21 13:57:29 CZ 1 NUSA PUTERA LOCH MELFORT 525005357 372924000 Roro / Passenger ferry Bulk carrier 4 2 209
12 01-Jul-21 14:21:25 CZ 4 MITRA NUSANTARA SC ETERNITY XLVII 525002063 525009331 Roro / Passenger ferry Gas tanker 3 5 33
13 01-Jul-21 14:52:51 CZ 4 KIRANA II MV GIAT 525015377 525019643 Roro / Passenger ferry General cargo 3 5
14 01-Jul-21 15:11:31 CZ 4 PORTLINK V BRAHMA 6 525005232 525024237 Roro / Passenger ferry Support vessel 3 6
15 01-Jul-21 18:19:25 CZ 2 TRIMAS FHADILA NSU NEWSTAR 525100688 353349000 Roro / Passenger ferry General cargo 3 1 40
16 02-Jul-21 06:25:35 CZ 1 SINARAN EMBUN ZHONG HAI CHANG YUN 6 563035330 414712000 Support vessel Bulk carrier 7 2 206
17 02-Jul-21 06:45:19 CZ 5 ZOEY DEBBIE 525200374 477217100 Roro / Passenger ferry Gas tanker 3 6 220
18 02-Jul-21 08:15:30 CZ 4 HAI ZHUANG 2 SINAR MASELA 525500603 525109005 Roro / Passenger ferry Oil products tanker 3 6 276 182
19 02-Jul-21 10:14:08 CZ 5 VIRGO 18 ANUGERAH DEWI 12 525006222 525020346 Roro / Passenger ferry Oil products tanker 4 5 106
20 02-Jul-21 13:14:55 CZ 5 FARINA NUSANTARA THOR MADOG 525002068 564794000 Roro / Passenger ferry General cargo 4 5 38
108
21 02-Jul-21 13:24:24 CZ 4 MUNIC 9 THOR MADOG 525200126 564794000 Roro / Passenger ferry General cargo 3 5 286 35
22 02-Jul-21 21:54:50 CZ 4 KMP DOROTHY GOLDSTAR SHINE 525100695 441382000 Roro / Passenger ferry Oil products tanker 3 6 177
23 03-Jul-21 06:31:31 CZ 4 KUMALA SEROJA XIII 525015375 525020229 Roro / Passenger ferry Oil products tanker 3 5
24 03-Jul-21 15:38:13 CZ 5 TITIAN MURNI KM SIN TOBA JAYA 68 525002061 525900986 Roro / Passenger ferry Fishing vessel 4 5 40
25 03-Jul-21 16:16:32 CZ 4 KM MUSTIKA KENCANA ANGELIA XVI 525015381 525016009 Roro / Passenger ferry Oil products tanker 3 5
26 03-Jul-21 23:16:04 CZ 1 KM MUSTIKA KENCANA MT ALMIRA XXII 525015381 525019324 Roro / Passenger ferry Oil products tanker 3 2 311
27 04-Jul-21 01:36:30 CZ 1 KMP LEGUNDI STB 36 525001125 525100325 Roro / Passenger ferry General cargo 3 2 280
28 04-Jul-21 14:34:00 CZ 4 SHALEM JAZEEL 525022100 525300489 Roro / Passenger ferry Oil products tanker 3 6 174
29 04-Jul-21 15:55:20 CZ 5 TRIMAS LAILA TIRTASARI 525018455 525007028 Roro / Passenger ferry Oil products tanker 4 5 28
30 04-Jul-21 19:20:20 CZ 1 KMP. HM BARUNA I DRAGON SKY 525004089 372572000 Roro / Passenger ferry General cargo 4 1 41
31 05-Jul-21 01:58:13 CZ 4 KMP LEGUNDI ZITA SCHULTE 525001125 235116958 Roro / Passenger ferry Oil products tanker 3 6 286 187
32 05-Jul-21 09:56:17 CZ 4 WIRA BERLIAN RUBY INDAH 525300391 564748000 Roro / Passenger ferry Bulk carrier 4 5 151 35
33 05-Jul-21 21:19:40 CZ 4 ATHAYA MATTINA 525119018 538008047 Roro / Passenger ferry Container 3 6 303 212
34 06-Jul-21 06:58:16 CZ 4 ELYSIA ZITA SCHULTE 525006235 235116958 Roro / Passenger ferry Oil products tanker 3 5 34
35 06-Jul-21 08:11:20 CZ 5 MUNIC 1 LEBAM 525005099 525002013 Roro / Passenger ferry General cargo 4 6
36 06-Jul-21 13:12:53 CZ 1 WIRA BERLIAN DL CARNATION 525300391 357909000 Roro / Passenger ferry Bulk carrier 4 1 107 27
37 06-Jul-21 21:11:33 CZ 5 SAFIRA NUSANTARA MT. JOHN CAINE 525002086 525019550 Roro / Passenger ferry Oil products tanker 4 5

38 07-Jul-21 04:17:23 CZ 2 KMP DOROTHY TINA 525100695 572939210 Roro / Passenger ferry General cargo 3 2 202
39 07-Jul-21 05:56:38 CZ 4 KUMALA SP 2 BSI 525015375 525022304 Roro / Passenger ferry Chemical tanker 4 6 184
40 07-Jul-21 12:13:20 CZ 1 KMP DOROTHY OOCL TEXAS 525100695 477036600 Roro / Passenger ferry Container 4 1 38
41 07-Jul-21 15:10:30 CZ 4 ATHAYA SINAR TARAKAN 525119018 525010240 Roro / Passenger ferry Gas tanker 4 5 75 6
42 07-Jul-21 15:57:19 CZ 4 KMP BSP 1 DIMAS PUTRA V 525012215 525021359 Roro / Passenger ferry Oil products tanker 3 6
Taking action
Crossing COG of COG of Avg. speed Avg. speed Role of Role of
Action taken based on
num. Vessel A Vessel B of Vessel A of Vessel B Vessel A Vessel B
COLREG?
1 90 0 7.24 6.38 Give-way Stand-on Vessel B give-way and changed the heading to the left to avoid crossing ahead Vessel A No
2 101 211 8.98 15.94 Stand-on Give-way Vessel A give-way and changed the heading to the left to avoid crossing ahead Vessel B No
4 100 31 13.92 4.82 Give-way Stand-on Vessel A give-way and changed the heading to the left to avoid crossing ahead Vessel B No
6 317 34 9.28 8.92 Stand-on Give-way No significant change is observed No
10 102 207 8.48 4.68 Stand-on Give-way Vessel B give-way and changed the heading to the right to avoid crossing ahead Vessel A Yes
109
24 107 40 8.45 4.86 Give-way Stand-on No significant change is observed No
25 289 23 6.00 7.60 Stand-on Give-way Vessel A give-way and changed the heading to the right to avoid crossing ahead Vessel B Yes

Appendix B
Process of working with IALA (International Association of Marine Aids to
Navigation and Lighthouse Authorities) Waterway Risk Assessment Program
(IWRAP)
Step 1) Setting the boundary
Set the boundary in the IWRAP by using model area tool. The boundary is used to
limit the analysis only to be in that area. In this study, the area is set as a rectangle
that bounded by four points as follows: P1 (-5.78018, 105.631), P2 (-5.78018,
106.071), P3 (-6.07616, 106.071), and P3 (-6.07616, 105.631).
Figure B1 Setting the boundary in IWRAP.
110
Step 2) Import AIS data.
The csv (comma-separated value) file of AIS data is imported into the IWRAP by
defining what data contained in that column. For example, Figure B2 is the sample
AIS data in the form of csv on July 2nd, 2021 and Figure B3 is the importing process.
Figure B2 The csv file of AIS data on July 2nd, 2021.
111
Figure B3 The importing process of AIS data.
Step 3) Extract the vessels’ trips.
After importing the AIS data, the trips are extracted by specifying the minimum
duration, minimum speed, minimum time below speed limit before stop, minimum
distance, and timeout. Those variables are used to calculate the number of trips in
IWRAP. If all values of variables are above the set value, than the trip is still
considered as a same trip.
Figure B3 The importing process of AIS data.
112
Step 4) Visual validation of IWRAP results.
As the behaviour of the IWRAP that determines a trip based on variables mentioned
in the previous step, a validation was performed to make sure the number of trips is
valid. The step is done by selecting each vessel name and check if the trip it has
made is the same as what it seems.
Figure B4 Visual validation of the trip.
Step 5) Set the DCPA value.
Figure B5 Setting the DCPA value in IWRAP.
113
To be able to get the near-miss crossing situation, we set the DCPA value as 1.0 nm
or equals to 1852 meters. The IWRAP shows an indicator for ships having the
DCPA less than or equals to the setting value.
Step 6) Acknowledge the near-miss situation.
The IWRAP shows an indicator for ships having the DCPA less than or equals to
the setting value.
Figure B5 An example of near-miss situation as indicated by IWRAP.
114
Appendix C
Variation Tree Analysis
C1. Mutiara Persada I (roro ferry) with Port Link III (roro ferry).
The incident happened on July 5th, 2017 in Port of Bakauheni.
KM Mutiara
Persada
KM Port Link III
115
C2. KM Dian No. 1 (roro ferry) with MT Soechi Chemical XIX (cargo vessel).
The collision accident happened on May 19th, 2010.
MT Soechi
Chemicals XIX
KM Dian No. 1
116
Appendix D
Preliminary Survey in Sunda Strait, December 7-8th, 2018.
A preliminary survey was conducted in Sunda Strait for two days to hear from
stakeholders involved in the operational of TSS Sunda Strait. This survey was
attended by Prof. Takeshi Shinoda, Fadilla Indrayuni Prastyasari, Dr. Putu Hangga
from Kyushu University and Dr. Dinariyana, and Dr. Buana from ITS Surabaya.
During this survey, the team visited ASDP, a state-owned company that operating
RoRo ferries; STC or ship traffic control that regulates the traffic of RoRo ferries;
BPTD an office that responsible to the scheduling and traffic monitoring of RoRo
ferries.
Day 1: Visit to PT ASDP Indonesia Ferry (Persero) (Angkutan Sungai, Danau

dan Penyeberangan / Passenger ferry operator in Indonesia) and BPTD (Balai
Pengelola Transportasi Darat / Land Transportation Management Office)
Banten
PT. ASDP Indonesia Ferry is a state-owned company that has a business in the ferry
operation in Indonesia. The total employee working in the ASDP Merak
approximately is 250 people. In the Sunda Strait, a route for passenger ferry is
operated by ASDP. This route is from Merak Port in West Java to the Bakauheni
Port in Lampung. The total number of vessels serving this route is 71, but 3 of them
cannot be operated due to major damage. Only 25 ferries are being operated
everyday by ASDP, because the berth available is only 5 and the load factor of each
ship is still low. Berths operated in the Merak are 5-1-2-3-6 while in Bakauheni are
7-1-2-3-6. The ship below 5000 GT has been excluded from the operation of Merak
117
Figure D-1 Discussion with ASDP employees about TSS.
Bakauheni, to reduce the number of the passenger ferry that is not operated.
The schedule of passenger ferry is regulated by BPTD (Balai Pengelola

Transportasi Darat) by considering the number of vessel and the vessel that can be
operated in one day. The result from the scheduling process decides that 1 ship can
be operated in 3 days and shall wait for the next trip in 6 days later. In each trip, a
passenger ferry needs 72 minutes to do loading and unloading process and 108
minutes for sailing. Hence, the total time for 1 trip is 3 hours. One ferry sails for 4
round-trips per day.
Ship Traffic Control (STC) is a body who controls the ferry when it approaches the
port. While the ship passing the Sunda strait is under the control of Vessel Traffic
Service (VTS). The STC determines whether the passenger ferry can enter and
leave the port or shall wait in the waiting zone. This waiting zone apparently is in
the precautionary area of the proposed TSS. This area is chosen as the waiting zone
118
Figure D-2 Discussion with BPTD about TSS.
due to the wind and current that is not too rough. STC also gives command to which
ship that can berth or unberth and in which berth she shall berth.
There is no incident or accident reported in Sunda Strait this year. Though, if a near-
miss condition happens, the officer of watch (OOW) in the ship will have a
communication with the OOW on the other ship. According to the ASDP, some of
the international cargo ship does not give any reply to the ferry when they try to call
them. Furthermore, the lack of English communication skill also gives a
contribution to the collision accident.
The impression of the ASDP of the proposed TSS is positive, but they gave some
suggestions to the team about the proposed TSS from their point of view.
• The ASDP questioned the use precautionary area in TSS and how is the
mechanism of the passenger ferry that crosses the Sunda Strait if there is
also a cargo ship passing the (Indonesian Archipelagic Sea State) IASL 1.
• The ferry is bounded by a schedule; hence it cannot wait for a very long
time (if the VTS tells the ferry to wait for the passing vessel).
119
• VTS shall control the traffic in the Sunda Strait (outside the authority of the
STC), hence the navigational safety can be increased.
Contact person:
ASDP
1. Diah Retno W (Ship Traffic Control)
2. Abdillah (Fleet manager)
3. Rudy Mahmudi (Port manager)
BPTD Banten
1. Nurhadi Unggul Wibowo (Head)
2. Eko (Operation manager)
Day 2: Sailing across Sunda Strait with KM. Sebuku
KM Sebuku has 30 crews onboard and led by a Captain. The crew of the ferry will
change shift every 12 hours (2 round trips). During the maneuvering period, there
are some officers ready on the bridge. One Captain, one officer of watch, giving
commands to the chief engineer who operates the control panel of the engine and
bow thruster, and one quarter master helps the OOW watching from the bridge. In
contrast, while the ship is in sailing condition, only one OOW and one quarter
master is working on the bridge.
During the trip from Bakauheni to Merak, KM Sebuku is in head-on condition with
KM Mufidah. The crew on the bridge made a communication with the crew onboard
KM Mufidah and decided to meet on the right-right.
120
Figure D-3 KM Sebuku on head-on condition with KM Mufidah.
Figure D-4 KM Situation on the bridge of KM Sebuku.
During the interview, the ships’ crew said that the main cause of the collision is
because human error. The lack of communication becomes the first cause because
the seafarers do not familiar with the maritime English. Hence, when the ship is on
an emergency condition the seafarers cannot make a communication properly that
leads on the collision accident.
121

Fadilla Dissertation

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Fadilla Dissertation

Uploaded by

Copyright:

Available Formats

PAGE 36 CONTAIN AIS DATA PROCESSING

SAFETY ASSESSMENT FRAMEWORK OF

Graduate School of Engineering

Keywords: AIS, Safety Assessment, Traffic Separation Scheme, Risk of Collision,

Firstly, I am profoundly indebted to my supervisor Prof. Takeshi Shinoda for his

Fadilla Indrayuni Prastyasari

Fukuoka, December 2021

1. Develop a framework for a vessel near-miss crossing situation database in

The implementation of TSS shall make a positive impact to the safety of

5. Assess the cost effectivity of TSS as a risk control option.

6. Give recommendations to the Government of Indonesia to improve the safety

As the final step of the safety assessment framework, recommendations are

1.3 Research contributions

This background of this research is due to the implementation of the traffic

1. Introducing a concept of vessel trip database and vessel near-miss crossing

1.4 List of Publications

Those contributions explained in the previous part is discussed and written in

1. Fadilla Indrayuni Prastyasari and Takeshi Shinoda (2019): Near-miss Analysis

1.5 Outline of the thesis

Chapter 2 provides a comprehensive literature study about the state-of-the-art of

Figure 1.1 Flow of dissertation (part 1).

As previously explained, the encountering situation is the initial stage of a collision

have a responsibility to supervising the traffic in ports, inland waterways, or coastal

Despite of the importance of risk assessment, neither The International Regulations

2.2 Methodologies for Probabilistic Risk Assessment

2.2.1 Ship Domain

A further development of the ship domain conducted to be applied in restricted

Figure 2.3 Geometrical collision diameter(Fujii and Shiobara, 1971).

Figure 2.4 Ship domains for head-on encounter (left)

Figure 2.5 The definition of CPA during a crossing encounter.

underestimate the risk of collision, which is undesired in this situation.

2.2.3 IALA Waterway Risk Assessment Program (IWRAP)

2.3 Safety Assessment Framework

Figure 2.8 A crossing situation happened after TSS (2).

Based on the methodology, safety assessment can be categorised into qualitative

3.1 Traffic Pattern in Sunda Strait

Figure 3.1 Location of archipelagic sea lanes in Indonesia (Dinariyana, 2020).

The characteristics of each course in Sunda Strait is explained as follows.

Figure 3.2 Traffic pattern in Sunda Strait before TSS.

Table 3-1 Example of vessels that navigating in Sunda Strait

Course Type of Vessel Length in Port of Origin Port of Destination

3.2 Environmental Considerations

to the Decree of the Minister of Forestry of the Republic of Indonesia No. 55 of

3.3 Accidents Record

3.4 AIS Messages

Automatic identification system (AIS) is a broadcasting system that enable a vessel

No. Date Ships Involved Location Losses

Back in 1989, a grounding accident was experienced by an oil tanker “Exxon

Figure 3.5 Illustration of AIS.

Despite some advantages offered by AIS, several considerations should be given to

Features Class A Class B

Static Data Dynamic Data Voyage-related Safety-related

3.5 AIS Data Ingestion

3.6 AIS Data Processing

3.7 Vessel Trips Database

Figure 3.14 Visual observation for the trip count

Figure 3.15 The number of trips that follow each course.

Figure 3.16 The number of trips made by each vessel type.

Table 3-5 Example of the near-miss database

4.2 Probability Assessment

The frequency, 𝑓, is affected by the number of vessels in conflict or in another term

An approach on how to do the probabilistic assessment is attempted, which is based